derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
6592 lines
215 KiB
LLVM
6592 lines
215 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=riscv32 -mattr=+f -verify-machineinstrs < %s \
|
|
; RUN: -target-abi=ilp32f | FileCheck -check-prefixes=RV32,RV32IF %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+f -verify-machineinstrs < %s \
|
|
; RUN: -target-abi=lp64f | FileCheck -check-prefixes=RV64,RV64IF %s
|
|
; RUN: llc -mtriple=riscv32 -mattr=+f,+d -verify-machineinstrs < %s \
|
|
; RUN: -target-abi=ilp32d | FileCheck -check-prefixes=RV32,RV32IFD %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+f,+d -verify-machineinstrs < %s \
|
|
; RUN: -target-abi=lp64d | FileCheck -check-prefixes=RV64,RV64IFD %s
|
|
|
|
; i32 saturate
|
|
|
|
define i32 @stest_f64i32(double %x) {
|
|
; RV32IF-LABEL: stest_f64i32:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfdi
|
|
; RV32IF-NEXT: lui a2, 524288
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: beqz a1, .LBB0_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: srli a4, a1, 31
|
|
; RV32IF-NEXT: j .LBB0_3
|
|
; RV32IF-NEXT: .LBB0_2:
|
|
; RV32IF-NEXT: sltu a4, a0, a3
|
|
; RV32IF-NEXT: .LBB0_3: # %entry
|
|
; RV32IF-NEXT: neg a5, a4
|
|
; RV32IF-NEXT: and a1, a5, a1
|
|
; RV32IF-NEXT: bnez a4, .LBB0_5
|
|
; RV32IF-NEXT: # %bb.4: # %entry
|
|
; RV32IF-NEXT: mv a0, a3
|
|
; RV32IF-NEXT: .LBB0_5: # %entry
|
|
; RV32IF-NEXT: li a3, -1
|
|
; RV32IF-NEXT: beq a1, a3, .LBB0_7
|
|
; RV32IF-NEXT: # %bb.6: # %entry
|
|
; RV32IF-NEXT: srli a1, a1, 31
|
|
; RV32IF-NEXT: xori a1, a1, 1
|
|
; RV32IF-NEXT: beqz a1, .LBB0_8
|
|
; RV32IF-NEXT: j .LBB0_9
|
|
; RV32IF-NEXT: .LBB0_7:
|
|
; RV32IF-NEXT: sltu a1, a2, a0
|
|
; RV32IF-NEXT: bnez a1, .LBB0_9
|
|
; RV32IF-NEXT: .LBB0_8: # %entry
|
|
; RV32IF-NEXT: lui a0, 524288
|
|
; RV32IF-NEXT: .LBB0_9: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i32:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfdi
|
|
; RV64IF-NEXT: lui a1, 524288
|
|
; RV64IF-NEXT: addiw a2, a1, -1
|
|
; RV64IF-NEXT: blt a0, a2, .LBB0_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a2
|
|
; RV64IF-NEXT: .LBB0_2: # %entry
|
|
; RV64IF-NEXT: blt a1, a0, .LBB0_4
|
|
; RV64IF-NEXT: # %bb.3: # %entry
|
|
; RV64IF-NEXT: lui a0, 524288
|
|
; RV64IF-NEXT: .LBB0_4: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i32:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i32:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 524288
|
|
; RV64IFD-NEXT: addiw a2, a1, -1
|
|
; RV64IFD-NEXT: bge a0, a2, .LBB0_3
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: bge a1, a0, .LBB0_4
|
|
; RV64IFD-NEXT: .LBB0_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
; RV64IFD-NEXT: .LBB0_3: # %entry
|
|
; RV64IFD-NEXT: mv a0, a2
|
|
; RV64IFD-NEXT: blt a1, a2, .LBB0_2
|
|
; RV64IFD-NEXT: .LBB0_4: # %entry
|
|
; RV64IFD-NEXT: lui a0, 524288
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i64
|
|
%0 = icmp slt i64 %conv, 2147483647
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 2147483647
|
|
%1 = icmp sgt i64 %spec.store.select, -2147483648
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 -2147483648
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utest_f64i32(double %x) {
|
|
; RV32IF-LABEL: utest_f64i32:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixunsdfdi
|
|
; RV32IF-NEXT: sltiu a2, a0, -1
|
|
; RV32IF-NEXT: seqz a1, a1
|
|
; RV32IF-NEXT: and a1, a1, a2
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: or a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: utest_f64i32:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixunsdfdi
|
|
; RV64IF-NEXT: li a1, -1
|
|
; RV64IF-NEXT: srli a1, a1, 32
|
|
; RV64IF-NEXT: bltu a0, a1, .LBB1_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB1_2: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i32:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: utest_f64i32:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.lu.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: li a1, -1
|
|
; RV64IFD-NEXT: srli a1, a1, 32
|
|
; RV64IFD-NEXT: bltu a0, a1, .LBB1_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB1_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i64
|
|
%0 = icmp ult i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f64i32(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i32:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfdi
|
|
; RV32IF-NEXT: beqz a1, .LBB2_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: srli a2, a1, 31
|
|
; RV32IF-NEXT: j .LBB2_3
|
|
; RV32IF-NEXT: .LBB2_2:
|
|
; RV32IF-NEXT: sltiu a2, a0, -1
|
|
; RV32IF-NEXT: .LBB2_3: # %entry
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: neg a2, a2
|
|
; RV32IF-NEXT: and a1, a2, a1
|
|
; RV32IF-NEXT: or a0, a3, a0
|
|
; RV32IF-NEXT: beqz a1, .LBB2_5
|
|
; RV32IF-NEXT: # %bb.4: # %entry
|
|
; RV32IF-NEXT: sgtz a1, a1
|
|
; RV32IF-NEXT: j .LBB2_6
|
|
; RV32IF-NEXT: .LBB2_5:
|
|
; RV32IF-NEXT: snez a1, a0
|
|
; RV32IF-NEXT: .LBB2_6: # %entry
|
|
; RV32IF-NEXT: neg a1, a1
|
|
; RV32IF-NEXT: and a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: ustest_f64i32:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfdi
|
|
; RV64IF-NEXT: li a1, -1
|
|
; RV64IF-NEXT: srli a1, a1, 32
|
|
; RV64IF-NEXT: blt a0, a1, .LBB2_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB2_2: # %entry
|
|
; RV64IF-NEXT: sgtz a1, a0
|
|
; RV64IF-NEXT: neg a1, a1
|
|
; RV64IF-NEXT: and a0, a1, a0
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i32:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: ustest_f64i32:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: li a1, -1
|
|
; RV64IFD-NEXT: srli a1, a1, 32
|
|
; RV64IFD-NEXT: blt a0, a1, .LBB2_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB2_2: # %entry
|
|
; RV64IFD-NEXT: sgtz a1, a0
|
|
; RV64IFD-NEXT: neg a1, a1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i64
|
|
%0 = icmp slt i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%1 = icmp sgt i64 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 0
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @stest_f32i32(float %x) {
|
|
; RV32-LABEL: stest_f32i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 524288
|
|
; RV64-NEXT: addiw a2, a1, -1
|
|
; RV64-NEXT: bge a0, a2, .LBB3_3
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: bge a1, a0, .LBB3_4
|
|
; RV64-NEXT: .LBB3_2: # %entry
|
|
; RV64-NEXT: ret
|
|
; RV64-NEXT: .LBB3_3: # %entry
|
|
; RV64-NEXT: mv a0, a2
|
|
; RV64-NEXT: blt a1, a2, .LBB3_2
|
|
; RV64-NEXT: .LBB3_4: # %entry
|
|
; RV64-NEXT: lui a0, 524288
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i64
|
|
%0 = icmp slt i64 %conv, 2147483647
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 2147483647
|
|
%1 = icmp sgt i64 %spec.store.select, -2147483648
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 -2147483648
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utest_f32i32(float %x) {
|
|
; RV32-LABEL: utest_f32i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: bltu a0, a1, .LBB4_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB4_2: # %entry
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i64
|
|
%0 = icmp ult i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f32i32(float %x) {
|
|
; RV32-LABEL: ustest_f32i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: blt a0, a1, .LBB5_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB5_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i64
|
|
%0 = icmp slt i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%1 = icmp sgt i64 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 0
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @stest_f16i32(half %x) {
|
|
; RV32-LABEL: stest_f16i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixsfdi
|
|
; RV32-NEXT: lui a2, 524288
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: beqz a1, .LBB6_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: srli a4, a1, 31
|
|
; RV32-NEXT: j .LBB6_3
|
|
; RV32-NEXT: .LBB6_2:
|
|
; RV32-NEXT: sltu a4, a0, a3
|
|
; RV32-NEXT: .LBB6_3: # %entry
|
|
; RV32-NEXT: neg a5, a4
|
|
; RV32-NEXT: and a1, a5, a1
|
|
; RV32-NEXT: bnez a4, .LBB6_5
|
|
; RV32-NEXT: # %bb.4: # %entry
|
|
; RV32-NEXT: mv a0, a3
|
|
; RV32-NEXT: .LBB6_5: # %entry
|
|
; RV32-NEXT: li a3, -1
|
|
; RV32-NEXT: beq a1, a3, .LBB6_7
|
|
; RV32-NEXT: # %bb.6: # %entry
|
|
; RV32-NEXT: srli a1, a1, 31
|
|
; RV32-NEXT: xori a1, a1, 1
|
|
; RV32-NEXT: beqz a1, .LBB6_8
|
|
; RV32-NEXT: j .LBB6_9
|
|
; RV32-NEXT: .LBB6_7:
|
|
; RV32-NEXT: sltu a1, a2, a0
|
|
; RV32-NEXT: bnez a1, .LBB6_9
|
|
; RV32-NEXT: .LBB6_8: # %entry
|
|
; RV32-NEXT: lui a0, 524288
|
|
; RV32-NEXT: .LBB6_9: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 524288
|
|
; RV64-NEXT: addiw a2, a1, -1
|
|
; RV64-NEXT: blt a0, a2, .LBB6_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a2
|
|
; RV64-NEXT: .LBB6_2: # %entry
|
|
; RV64-NEXT: blt a1, a0, .LBB6_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: lui a0, 524288
|
|
; RV64-NEXT: .LBB6_4: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i64
|
|
%0 = icmp slt i64 %conv, 2147483647
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 2147483647
|
|
%1 = icmp sgt i64 %spec.store.select, -2147483648
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 -2147483648
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utest_f16i32(half %x) {
|
|
; RV32-LABEL: utest_f16i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixunssfdi
|
|
; RV32-NEXT: sltiu a2, a0, -1
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: and a1, a1, a2
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: or a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f16i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: bltu a0, a1, .LBB7_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB7_2: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i64
|
|
%0 = icmp ult i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f16i32(half %x) {
|
|
; RV32-LABEL: ustest_f16i32:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixsfdi
|
|
; RV32-NEXT: beqz a1, .LBB8_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: j .LBB8_3
|
|
; RV32-NEXT: .LBB8_2:
|
|
; RV32-NEXT: sltiu a2, a0, -1
|
|
; RV32-NEXT: .LBB8_3: # %entry
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: neg a2, a2
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: or a0, a3, a0
|
|
; RV32-NEXT: beqz a1, .LBB8_5
|
|
; RV32-NEXT: # %bb.4: # %entry
|
|
; RV32-NEXT: sgtz a1, a1
|
|
; RV32-NEXT: j .LBB8_6
|
|
; RV32-NEXT: .LBB8_5:
|
|
; RV32-NEXT: snez a1, a0
|
|
; RV32-NEXT: .LBB8_6: # %entry
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i32:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: blt a0, a1, .LBB8_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB8_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i64
|
|
%0 = icmp slt i64 %conv, 4294967295
|
|
%spec.store.select = select i1 %0, i64 %conv, i64 4294967295
|
|
%1 = icmp sgt i64 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i64 %spec.store.select, i64 0
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
; i16 saturate
|
|
|
|
define i16 @stest_f64i16(double %x) {
|
|
; RV32IF-LABEL: stest_f64i16:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfsi
|
|
; RV32IF-NEXT: lui a1, 8
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: blt a0, a1, .LBB9_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB9_2: # %entry
|
|
; RV32IF-NEXT: lui a1, 1048568
|
|
; RV32IF-NEXT: blt a1, a0, .LBB9_4
|
|
; RV32IF-NEXT: # %bb.3: # %entry
|
|
; RV32IF-NEXT: lui a0, 1048568
|
|
; RV32IF-NEXT: .LBB9_4: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i16:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfsi
|
|
; RV64IF-NEXT: lui a1, 8
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: blt a0, a1, .LBB9_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB9_2: # %entry
|
|
; RV64IF-NEXT: lui a1, 1048568
|
|
; RV64IF-NEXT: blt a1, a0, .LBB9_4
|
|
; RV64IF-NEXT: # %bb.3: # %entry
|
|
; RV64IF-NEXT: lui a0, 1048568
|
|
; RV64IF-NEXT: .LBB9_4: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i16:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 8
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: bge a0, a1, .LBB9_3
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: lui a1, 1048568
|
|
; RV32IFD-NEXT: bge a1, a0, .LBB9_4
|
|
; RV32IFD-NEXT: .LBB9_2: # %entry
|
|
; RV32IFD-NEXT: ret
|
|
; RV32IFD-NEXT: .LBB9_3: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: lui a1, 1048568
|
|
; RV32IFD-NEXT: blt a1, a0, .LBB9_2
|
|
; RV32IFD-NEXT: .LBB9_4: # %entry
|
|
; RV32IFD-NEXT: lui a0, 1048568
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i16:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 8
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: bge a0, a1, .LBB9_3
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: lui a1, 1048568
|
|
; RV64IFD-NEXT: bge a1, a0, .LBB9_4
|
|
; RV64IFD-NEXT: .LBB9_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
; RV64IFD-NEXT: .LBB9_3: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: lui a1, 1048568
|
|
; RV64IFD-NEXT: blt a1, a0, .LBB9_2
|
|
; RV64IFD-NEXT: .LBB9_4: # %entry
|
|
; RV64IFD-NEXT: lui a0, 1048568
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i32
|
|
%0 = icmp slt i32 %conv, 32767
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 32767
|
|
%1 = icmp sgt i32 %spec.store.select, -32768
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 -32768
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utest_f64i16(double %x) {
|
|
; RV32IF-LABEL: utest_f64i16:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixunsdfsi
|
|
; RV32IF-NEXT: lui a1, 16
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: bltu a0, a1, .LBB10_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB10_2: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: utest_f64i16:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixunsdfsi
|
|
; RV64IF-NEXT: lui a1, 16
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: bltu a0, a1, .LBB10_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB10_2: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i16:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 16
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: bltu a0, a1, .LBB10_2
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: .LBB10_2: # %entry
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: utest_f64i16:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 16
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: bltu a0, a1, .LBB10_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB10_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i32
|
|
%0 = icmp ult i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f64i16(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i16:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfsi
|
|
; RV32IF-NEXT: lui a1, 16
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: blt a0, a1, .LBB11_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB11_2: # %entry
|
|
; RV32IF-NEXT: sgtz a1, a0
|
|
; RV32IF-NEXT: neg a1, a1
|
|
; RV32IF-NEXT: and a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: ustest_f64i16:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfsi
|
|
; RV64IF-NEXT: lui a1, 16
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: blt a0, a1, .LBB11_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB11_2: # %entry
|
|
; RV64IF-NEXT: sgtz a1, a0
|
|
; RV64IF-NEXT: neg a1, a1
|
|
; RV64IF-NEXT: and a0, a1, a0
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i16:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 16
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: blt a0, a1, .LBB11_2
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: .LBB11_2: # %entry
|
|
; RV32IFD-NEXT: sgtz a1, a0
|
|
; RV32IFD-NEXT: neg a1, a1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: ustest_f64i16:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 16
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: blt a0, a1, .LBB11_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB11_2: # %entry
|
|
; RV64IFD-NEXT: sgtz a1, a0
|
|
; RV64IFD-NEXT: neg a1, a1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i32
|
|
%0 = icmp slt i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%1 = icmp sgt i32 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 0
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @stest_f32i16(float %x) {
|
|
; RV32-LABEL: stest_f32i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 8
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bge a0, a1, .LBB12_3
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: bge a1, a0, .LBB12_4
|
|
; RV32-NEXT: .LBB12_2: # %entry
|
|
; RV32-NEXT: ret
|
|
; RV32-NEXT: .LBB12_3: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: blt a1, a0, .LBB12_2
|
|
; RV32-NEXT: .LBB12_4: # %entry
|
|
; RV32-NEXT: lui a0, 1048568
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 8
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bge a0, a1, .LBB12_3
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: bge a1, a0, .LBB12_4
|
|
; RV64-NEXT: .LBB12_2: # %entry
|
|
; RV64-NEXT: ret
|
|
; RV64-NEXT: .LBB12_3: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: blt a1, a0, .LBB12_2
|
|
; RV64-NEXT: .LBB12_4: # %entry
|
|
; RV64-NEXT: lui a0, 1048568
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i32
|
|
%0 = icmp slt i32 %conv, 32767
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 32767
|
|
%1 = icmp sgt i32 %spec.store.select, -32768
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 -32768
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utest_f32i16(float %x) {
|
|
; RV32-LABEL: utest_f32i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bltu a0, a1, .LBB13_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB13_2: # %entry
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bltu a0, a1, .LBB13_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB13_2: # %entry
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i32
|
|
%0 = icmp ult i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f32i16(float %x) {
|
|
; RV32-LABEL: ustest_f32i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB14_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB14_2: # %entry
|
|
; RV32-NEXT: sgtz a1, a0
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB14_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB14_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i32
|
|
%0 = icmp slt i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%1 = icmp sgt i32 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 0
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @stest_f16i16(half %x) {
|
|
; RV32-LABEL: stest_f16i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 8
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB15_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB15_2: # %entry
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: blt a1, a0, .LBB15_4
|
|
; RV32-NEXT: # %bb.3: # %entry
|
|
; RV32-NEXT: lui a0, 1048568
|
|
; RV32-NEXT: .LBB15_4: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 8
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB15_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB15_2: # %entry
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: blt a1, a0, .LBB15_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: lui a0, 1048568
|
|
; RV64-NEXT: .LBB15_4: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i32
|
|
%0 = icmp slt i32 %conv, 32767
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 32767
|
|
%1 = icmp sgt i32 %spec.store.select, -32768
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 -32768
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utest_f16i16(half %x) {
|
|
; RV32-LABEL: utest_f16i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bltu a0, a1, .LBB16_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB16_2: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f16i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bltu a0, a1, .LBB16_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB16_2: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i32
|
|
%0 = icmp ult i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f16i16(half %x) {
|
|
; RV32-LABEL: ustest_f16i16:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB17_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB17_2: # %entry
|
|
; RV32-NEXT: sgtz a1, a0
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i16:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB17_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB17_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i32
|
|
%0 = icmp slt i32 %conv, 65535
|
|
%spec.store.select = select i1 %0, i32 %conv, i32 65535
|
|
%1 = icmp sgt i32 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i32 %spec.store.select, i32 0
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
; i64 saturate
|
|
|
|
define i64 @stest_f64i64(double %x) {
|
|
; RV32IF-LABEL: stest_f64i64:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: addi sp, sp, -80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32IF-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s9, 36(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: .cfi_offset s0, -8
|
|
; RV32IF-NEXT: .cfi_offset s1, -12
|
|
; RV32IF-NEXT: .cfi_offset s2, -16
|
|
; RV32IF-NEXT: .cfi_offset s3, -20
|
|
; RV32IF-NEXT: .cfi_offset s4, -24
|
|
; RV32IF-NEXT: .cfi_offset s5, -28
|
|
; RV32IF-NEXT: .cfi_offset s6, -32
|
|
; RV32IF-NEXT: .cfi_offset s7, -36
|
|
; RV32IF-NEXT: .cfi_offset s8, -40
|
|
; RV32IF-NEXT: .cfi_offset s9, -44
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB18_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li s1, 0
|
|
; RV32IF-NEXT: li s2, 0
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: j .LBB18_8
|
|
; RV32IF-NEXT: .LBB18_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: srai s0, a1, 31
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: ori s2, s0, 1
|
|
; RV32IF-NEXT: addi a5, a3, -1
|
|
; RV32IF-NEXT: addi a3, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a5
|
|
; RV32IF-NEXT: add a0, a1, a4
|
|
; RV32IF-NEXT: seqz a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a5
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a0, a0, a1
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB18_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a1, 1043
|
|
; RV32IF-NEXT: li a4, 1075
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: sub a2, a4, a2
|
|
; RV32IF-NEXT: srai a4, a1, 31
|
|
; RV32IF-NEXT: srl a5, a0, a2
|
|
; RV32IF-NEXT: and s3, a4, a5
|
|
; RV32IF-NEXT: bltz a1, .LBB18_6
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl s4, a0, a1
|
|
; RV32IF-NEXT: j .LBB18_7
|
|
; RV32IF-NEXT: .LBB18_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: sw a3, 16(sp)
|
|
; RV32IF-NEXT: sw a0, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 0(a1)
|
|
; RV32IF-NEXT: lw a2, 4(a1)
|
|
; RV32IF-NEXT: lw a5, 8(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a6, a4, 1
|
|
; RV32IF-NEXT: srli a7, a2, 1
|
|
; RV32IF-NEXT: srli t0, a5, 1
|
|
; RV32IF-NEXT: srl a6, a6, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl a3, t0, a3
|
|
; RV32IF-NEXT: sll a2, a2, a0
|
|
; RV32IF-NEXT: sll a5, a5, a0
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or s3, a2, a6
|
|
; RV32IF-NEXT: or a2, a5, a7
|
|
; RV32IF-NEXT: or a3, a1, a3
|
|
; RV32IF-NEXT: sll s4, a4, a0
|
|
; RV32IF-NEXT: mv a0, s2
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s7, a0, s1
|
|
; RV32IF-NEXT: sltu a0, s7, a0
|
|
; RV32IF-NEXT: add s8, a1, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: add s2, s5, a1
|
|
; RV32IF-NEXT: sltu a0, s2, s5
|
|
; RV32IF-NEXT: add s6, s6, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s4, s6, a0
|
|
; RV32IF-NEXT: sltu s5, s4, s6
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s4, a0, s4
|
|
; RV32IF-NEXT: sltu a2, s4, a0
|
|
; RV32IF-NEXT: add a0, s4, s7
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: add a1, a1, s8
|
|
; RV32IF-NEXT: sltu a2, a0, s4
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: j .LBB18_8
|
|
; RV32IF-NEXT: .LBB18_6:
|
|
; RV32IF-NEXT: srl a1, a3, a2
|
|
; RV32IF-NEXT: not a2, a2
|
|
; RV32IF-NEXT: slli a0, a0, 1
|
|
; RV32IF-NEXT: sll a0, a0, a2
|
|
; RV32IF-NEXT: or s4, a1, a0
|
|
; RV32IF-NEXT: .LBB18_7: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add s7, a1, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s8, s7, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: add s9, a0, s8
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a0, s9, a0
|
|
; RV32IF-NEXT: sltu a2, s9, s5
|
|
; RV32IF-NEXT: sltu a3, s8, s7
|
|
; RV32IF-NEXT: sltu a4, a0, s9
|
|
; RV32IF-NEXT: add a3, s6, a3
|
|
; RV32IF-NEXT: add a2, a3, a2
|
|
; RV32IF-NEXT: add a1, a2, a1
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: .LBB18_8: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: lui a2, 524288
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: beq s2, a3, .LBB18_10
|
|
; RV32IF-NEXT: # %bb.9: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: sltu a4, s2, a3
|
|
; RV32IF-NEXT: or a5, a0, a1
|
|
; RV32IF-NEXT: bnez a5, .LBB18_11
|
|
; RV32IF-NEXT: j .LBB18_12
|
|
; RV32IF-NEXT: .LBB18_10:
|
|
; RV32IF-NEXT: sltiu a4, s1, -1
|
|
; RV32IF-NEXT: or a5, a0, a1
|
|
; RV32IF-NEXT: beqz a5, .LBB18_12
|
|
; RV32IF-NEXT: .LBB18_11: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a4, a1, 31
|
|
; RV32IF-NEXT: .LBB18_12: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: neg a5, a4
|
|
; RV32IF-NEXT: addi a6, a4, -1
|
|
; RV32IF-NEXT: bnez a4, .LBB18_14
|
|
; RV32IF-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: mv s2, a3
|
|
; RV32IF-NEXT: .LBB18_14: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: or a3, a6, s1
|
|
; RV32IF-NEXT: and a1, a5, a1
|
|
; RV32IF-NEXT: and a4, a5, a0
|
|
; RV32IF-NEXT: beq s2, a2, .LBB18_16
|
|
; RV32IF-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: sltu a0, a2, s2
|
|
; RV32IF-NEXT: j .LBB18_17
|
|
; RV32IF-NEXT: .LBB18_16:
|
|
; RV32IF-NEXT: snez a0, a3
|
|
; RV32IF-NEXT: .LBB18_17: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: and a4, a4, a1
|
|
; RV32IF-NEXT: li a2, -1
|
|
; RV32IF-NEXT: beq a4, a2, .LBB18_19
|
|
; RV32IF-NEXT: # %bb.18: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a1, a1, 31
|
|
; RV32IF-NEXT: xori a0, a1, 1
|
|
; RV32IF-NEXT: .LBB18_19: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: bnez a0, .LBB18_21
|
|
; RV32IF-NEXT: # %bb.20: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: lui s2, 524288
|
|
; RV32IF-NEXT: .LBB18_21: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: neg a0, a0
|
|
; RV32IF-NEXT: and a0, a0, a3
|
|
; RV32IF-NEXT: mv a1, s2
|
|
; RV32IF-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s9, 36(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: .cfi_restore s0
|
|
; RV32IF-NEXT: .cfi_restore s1
|
|
; RV32IF-NEXT: .cfi_restore s2
|
|
; RV32IF-NEXT: .cfi_restore s3
|
|
; RV32IF-NEXT: .cfi_restore s4
|
|
; RV32IF-NEXT: .cfi_restore s5
|
|
; RV32IF-NEXT: .cfi_restore s6
|
|
; RV32IF-NEXT: .cfi_restore s7
|
|
; RV32IF-NEXT: .cfi_restore s8
|
|
; RV32IF-NEXT: .cfi_restore s9
|
|
; RV32IF-NEXT: addi sp, sp, 80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i64:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfti
|
|
; RV64IF-NEXT: li a2, -1
|
|
; RV64IF-NEXT: srli a3, a2, 1
|
|
; RV64IF-NEXT: beqz a1, .LBB18_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: srli a4, a1, 63
|
|
; RV64IF-NEXT: j .LBB18_3
|
|
; RV64IF-NEXT: .LBB18_2:
|
|
; RV64IF-NEXT: sltu a4, a0, a3
|
|
; RV64IF-NEXT: .LBB18_3: # %entry
|
|
; RV64IF-NEXT: neg a5, a4
|
|
; RV64IF-NEXT: and a5, a5, a1
|
|
; RV64IF-NEXT: bnez a4, .LBB18_5
|
|
; RV64IF-NEXT: # %bb.4: # %entry
|
|
; RV64IF-NEXT: mv a0, a3
|
|
; RV64IF-NEXT: .LBB18_5: # %entry
|
|
; RV64IF-NEXT: slli a1, a2, 63
|
|
; RV64IF-NEXT: beq a5, a2, .LBB18_7
|
|
; RV64IF-NEXT: # %bb.6: # %entry
|
|
; RV64IF-NEXT: srli a5, a5, 63
|
|
; RV64IF-NEXT: xori a2, a5, 1
|
|
; RV64IF-NEXT: beqz a2, .LBB18_8
|
|
; RV64IF-NEXT: j .LBB18_9
|
|
; RV64IF-NEXT: .LBB18_7:
|
|
; RV64IF-NEXT: sltu a2, a1, a0
|
|
; RV64IF-NEXT: bnez a2, .LBB18_9
|
|
; RV64IF-NEXT: .LBB18_8: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB18_9: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i64:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 96
|
|
; RV32IFD-NEXT: sw ra, 92(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s0, 88(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s1, 84(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s2, 80(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s3, 76(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s4, 72(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s5, 68(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s6, 64(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s7, 60(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s8, 56(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s9, 52(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: .cfi_offset ra, -4
|
|
; RV32IFD-NEXT: .cfi_offset s0, -8
|
|
; RV32IFD-NEXT: .cfi_offset s1, -12
|
|
; RV32IFD-NEXT: .cfi_offset s2, -16
|
|
; RV32IFD-NEXT: .cfi_offset s3, -20
|
|
; RV32IFD-NEXT: .cfi_offset s4, -24
|
|
; RV32IFD-NEXT: .cfi_offset s5, -28
|
|
; RV32IFD-NEXT: .cfi_offset s6, -32
|
|
; RV32IFD-NEXT: .cfi_offset s7, -36
|
|
; RV32IFD-NEXT: .cfi_offset s8, -40
|
|
; RV32IFD-NEXT: .cfi_offset s9, -44
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB18_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li s1, 0
|
|
; RV32IFD-NEXT: li s2, 0
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: j .LBB18_8
|
|
; RV32IFD-NEXT: .LBB18_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: srai s0, a1, 31
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: ori s2, s0, 1
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB18_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 1043
|
|
; RV32IFD-NEXT: li a4, 1075
|
|
; RV32IFD-NEXT: sub a3, a3, a0
|
|
; RV32IFD-NEXT: sub a0, a4, a0
|
|
; RV32IFD-NEXT: srai a4, a3, 31
|
|
; RV32IFD-NEXT: srl a5, a1, a0
|
|
; RV32IFD-NEXT: and s3, a4, a5
|
|
; RV32IFD-NEXT: bltz a3, .LBB18_6
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl s4, a1, a3
|
|
; RV32IFD-NEXT: j .LBB18_7
|
|
; RV32IFD-NEXT: .LBB18_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 0(a1)
|
|
; RV32IFD-NEXT: lw a2, 4(a1)
|
|
; RV32IFD-NEXT: lw a5, 8(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a6, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a2, 1
|
|
; RV32IFD-NEXT: srli t0, a5, 1
|
|
; RV32IFD-NEXT: srl a6, a6, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl a3, t0, a3
|
|
; RV32IFD-NEXT: sll a2, a2, a0
|
|
; RV32IFD-NEXT: sll a5, a5, a0
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or s3, a2, a6
|
|
; RV32IFD-NEXT: or a2, a5, a7
|
|
; RV32IFD-NEXT: or a3, a1, a3
|
|
; RV32IFD-NEXT: sll s4, a4, a0
|
|
; RV32IFD-NEXT: mv a0, s2
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s7, a0, s1
|
|
; RV32IFD-NEXT: sltu a0, s7, a0
|
|
; RV32IFD-NEXT: add s8, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: add s2, s5, a1
|
|
; RV32IFD-NEXT: sltu a0, s2, s5
|
|
; RV32IFD-NEXT: add s6, s6, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s4, s6, a0
|
|
; RV32IFD-NEXT: sltu s5, s4, s6
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s4, a0, s4
|
|
; RV32IFD-NEXT: sltu a2, s4, a0
|
|
; RV32IFD-NEXT: add a0, s4, s7
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: add a1, a1, s8
|
|
; RV32IFD-NEXT: sltu a2, a0, s4
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: j .LBB18_8
|
|
; RV32IFD-NEXT: .LBB18_6:
|
|
; RV32IFD-NEXT: srl a2, a2, a0
|
|
; RV32IFD-NEXT: not a0, a0
|
|
; RV32IFD-NEXT: slli a1, a1, 1
|
|
; RV32IFD-NEXT: sll a0, a1, a0
|
|
; RV32IFD-NEXT: or s4, a2, a0
|
|
; RV32IFD-NEXT: .LBB18_7: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add s7, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s8, s7, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: add s9, a0, s8
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a0, s9, a0
|
|
; RV32IFD-NEXT: sltu a2, s9, s5
|
|
; RV32IFD-NEXT: sltu a3, s8, s7
|
|
; RV32IFD-NEXT: sltu a4, a0, s9
|
|
; RV32IFD-NEXT: add a3, s6, a3
|
|
; RV32IFD-NEXT: add a2, a3, a2
|
|
; RV32IFD-NEXT: add a1, a2, a1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: .LBB18_8: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: lui a2, 524288
|
|
; RV32IFD-NEXT: addi a3, a2, -1
|
|
; RV32IFD-NEXT: beq s2, a3, .LBB18_10
|
|
; RV32IFD-NEXT: # %bb.9: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: sltu a4, s2, a3
|
|
; RV32IFD-NEXT: or a5, a0, a1
|
|
; RV32IFD-NEXT: bnez a5, .LBB18_11
|
|
; RV32IFD-NEXT: j .LBB18_12
|
|
; RV32IFD-NEXT: .LBB18_10:
|
|
; RV32IFD-NEXT: sltiu a4, s1, -1
|
|
; RV32IFD-NEXT: or a5, a0, a1
|
|
; RV32IFD-NEXT: beqz a5, .LBB18_12
|
|
; RV32IFD-NEXT: .LBB18_11: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a4, a1, 31
|
|
; RV32IFD-NEXT: .LBB18_12: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: neg a5, a4
|
|
; RV32IFD-NEXT: addi a6, a4, -1
|
|
; RV32IFD-NEXT: bnez a4, .LBB18_14
|
|
; RV32IFD-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: mv s2, a3
|
|
; RV32IFD-NEXT: .LBB18_14: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: or a3, a6, s1
|
|
; RV32IFD-NEXT: and a1, a5, a1
|
|
; RV32IFD-NEXT: and a4, a5, a0
|
|
; RV32IFD-NEXT: beq s2, a2, .LBB18_16
|
|
; RV32IFD-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: sltu a0, a2, s2
|
|
; RV32IFD-NEXT: j .LBB18_17
|
|
; RV32IFD-NEXT: .LBB18_16:
|
|
; RV32IFD-NEXT: snez a0, a3
|
|
; RV32IFD-NEXT: .LBB18_17: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: and a4, a4, a1
|
|
; RV32IFD-NEXT: li a2, -1
|
|
; RV32IFD-NEXT: beq a4, a2, .LBB18_19
|
|
; RV32IFD-NEXT: # %bb.18: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a1, a1, 31
|
|
; RV32IFD-NEXT: xori a0, a1, 1
|
|
; RV32IFD-NEXT: .LBB18_19: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: bnez a0, .LBB18_21
|
|
; RV32IFD-NEXT: # %bb.20: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: lui s2, 524288
|
|
; RV32IFD-NEXT: .LBB18_21: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: neg a0, a0
|
|
; RV32IFD-NEXT: and a0, a0, a3
|
|
; RV32IFD-NEXT: mv a1, s2
|
|
; RV32IFD-NEXT: lw ra, 92(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s0, 88(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s1, 84(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s2, 80(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s3, 76(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s4, 72(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s5, 68(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s6, 64(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s7, 60(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s8, 56(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s9, 52(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: .cfi_restore ra
|
|
; RV32IFD-NEXT: .cfi_restore s0
|
|
; RV32IFD-NEXT: .cfi_restore s1
|
|
; RV32IFD-NEXT: .cfi_restore s2
|
|
; RV32IFD-NEXT: .cfi_restore s3
|
|
; RV32IFD-NEXT: .cfi_restore s4
|
|
; RV32IFD-NEXT: .cfi_restore s5
|
|
; RV32IFD-NEXT: .cfi_restore s6
|
|
; RV32IFD-NEXT: .cfi_restore s7
|
|
; RV32IFD-NEXT: .cfi_restore s8
|
|
; RV32IFD-NEXT: .cfi_restore s9
|
|
; RV32IFD-NEXT: addi sp, sp, 96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i64:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV64IFD-NEXT: seqz a1, a1
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i128
|
|
%0 = icmp slt i128 %conv, 9223372036854775807
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 9223372036854775807
|
|
%1 = icmp sgt i128 %spec.store.select, -9223372036854775808
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 -9223372036854775808
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utest_f64i64(double %x) {
|
|
; RV32IF-LABEL: utest_f64i64:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB19_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: li a2, 0
|
|
; RV32IF-NEXT: j .LBB19_8
|
|
; RV32IF-NEXT: .LBB19_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: addi a3, a3, -1
|
|
; RV32IF-NEXT: addi a0, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: seqz a4, a0
|
|
; RV32IF-NEXT: add a3, a4, a3
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a1, a1, a3
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB19_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a5, 1075
|
|
; RV32IF-NEXT: li a4, 1043
|
|
; RV32IF-NEXT: sub a4, a4, a2
|
|
; RV32IF-NEXT: sub a5, a5, a2
|
|
; RV32IF-NEXT: bltz a4, .LBB19_6
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl a0, a1, a4
|
|
; RV32IF-NEXT: j .LBB19_7
|
|
; RV32IF-NEXT: .LBB19_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: addi sp, sp, -32
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 32
|
|
; RV32IF-NEXT: sw a0, 16(sp)
|
|
; RV32IF-NEXT: sw a1, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 8(a1)
|
|
; RV32IF-NEXT: lw a5, 4(a1)
|
|
; RV32IF-NEXT: lw a6, 0(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a2, a4, 1
|
|
; RV32IF-NEXT: srli a7, a5, 1
|
|
; RV32IF-NEXT: srli t0, a6, 1
|
|
; RV32IF-NEXT: srl a2, a2, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl t0, t0, a3
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or a2, a1, a2
|
|
; RV32IF-NEXT: sll a1, a4, a0
|
|
; RV32IF-NEXT: sll a4, a5, a0
|
|
; RV32IF-NEXT: or a3, a1, a7
|
|
; RV32IF-NEXT: or a1, a4, t0
|
|
; RV32IF-NEXT: sll a0, a6, a0
|
|
; RV32IF-NEXT: addi sp, sp, 32
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: j .LBB19_8
|
|
; RV32IF-NEXT: .LBB19_6:
|
|
; RV32IF-NEXT: srl a0, a0, a5
|
|
; RV32IF-NEXT: not a2, a5
|
|
; RV32IF-NEXT: slli a3, a1, 1
|
|
; RV32IF-NEXT: sll a2, a3, a2
|
|
; RV32IF-NEXT: or a0, a0, a2
|
|
; RV32IF-NEXT: .LBB19_7: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: li a2, 0
|
|
; RV32IF-NEXT: srl a1, a1, a5
|
|
; RV32IF-NEXT: srai a4, a4, 31
|
|
; RV32IF-NEXT: and a1, a4, a1
|
|
; RV32IF-NEXT: .LBB19_8: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: or a4, a2, a3
|
|
; RV32IF-NEXT: xori a3, a3, 1
|
|
; RV32IF-NEXT: seqz a4, a4
|
|
; RV32IF-NEXT: or a2, a3, a2
|
|
; RV32IF-NEXT: seqz a2, a2
|
|
; RV32IF-NEXT: addi a2, a2, -1
|
|
; RV32IF-NEXT: and a2, a2, a4
|
|
; RV32IF-NEXT: neg a2, a2
|
|
; RV32IF-NEXT: and a0, a2, a0
|
|
; RV32IF-NEXT: and a1, a2, a1
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f64i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixunsdfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i64:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -48
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 48
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB19_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: li a2, 0
|
|
; RV32IFD-NEXT: j .LBB19_8
|
|
; RV32IFD-NEXT: .LBB19_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB19_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a5, 1075
|
|
; RV32IFD-NEXT: li a4, 1043
|
|
; RV32IFD-NEXT: sub a4, a4, a0
|
|
; RV32IFD-NEXT: sub a5, a5, a0
|
|
; RV32IFD-NEXT: bltz a4, .LBB19_6
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl a0, a1, a4
|
|
; RV32IFD-NEXT: j .LBB19_7
|
|
; RV32IFD-NEXT: .LBB19_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 8(a1)
|
|
; RV32IFD-NEXT: lw a5, 4(a1)
|
|
; RV32IFD-NEXT: lw a6, 0(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a2, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a5, 1
|
|
; RV32IFD-NEXT: srli t0, a6, 1
|
|
; RV32IFD-NEXT: srl a2, a2, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl t0, t0, a3
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or a2, a1, a2
|
|
; RV32IFD-NEXT: sll a1, a4, a0
|
|
; RV32IFD-NEXT: sll a4, a5, a0
|
|
; RV32IFD-NEXT: or a3, a1, a7
|
|
; RV32IFD-NEXT: or a1, a4, t0
|
|
; RV32IFD-NEXT: sll a0, a6, a0
|
|
; RV32IFD-NEXT: j .LBB19_8
|
|
; RV32IFD-NEXT: .LBB19_6:
|
|
; RV32IFD-NEXT: srl a0, a2, a5
|
|
; RV32IFD-NEXT: not a2, a5
|
|
; RV32IFD-NEXT: slli a3, a1, 1
|
|
; RV32IFD-NEXT: sll a2, a3, a2
|
|
; RV32IFD-NEXT: or a0, a0, a2
|
|
; RV32IFD-NEXT: .LBB19_7: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: li a2, 0
|
|
; RV32IFD-NEXT: srl a1, a1, a5
|
|
; RV32IFD-NEXT: srai a4, a4, 31
|
|
; RV32IFD-NEXT: and a1, a4, a1
|
|
; RV32IFD-NEXT: .LBB19_8: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: or a4, a2, a3
|
|
; RV32IFD-NEXT: xori a3, a3, 1
|
|
; RV32IFD-NEXT: seqz a4, a4
|
|
; RV32IFD-NEXT: or a2, a3, a2
|
|
; RV32IFD-NEXT: seqz a2, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: and a2, a2, a4
|
|
; RV32IFD-NEXT: neg a2, a2
|
|
; RV32IFD-NEXT: and a0, a2, a0
|
|
; RV32IFD-NEXT: and a1, a2, a1
|
|
; RV32IFD-NEXT: addi sp, sp, 48
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i128
|
|
%0 = icmp ult i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f64i64(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i64:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: addi sp, sp, -80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32IF-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s9, 36(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: .cfi_offset s0, -8
|
|
; RV32IF-NEXT: .cfi_offset s1, -12
|
|
; RV32IF-NEXT: .cfi_offset s2, -16
|
|
; RV32IF-NEXT: .cfi_offset s3, -20
|
|
; RV32IF-NEXT: .cfi_offset s4, -24
|
|
; RV32IF-NEXT: .cfi_offset s5, -28
|
|
; RV32IF-NEXT: .cfi_offset s6, -32
|
|
; RV32IF-NEXT: .cfi_offset s7, -36
|
|
; RV32IF-NEXT: .cfi_offset s8, -40
|
|
; RV32IF-NEXT: .cfi_offset s9, -44
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB20_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li s1, 0
|
|
; RV32IF-NEXT: li s6, 0
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: j .LBB20_6
|
|
; RV32IF-NEXT: .LBB20_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: srai s0, a1, 31
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: ori s2, s0, 1
|
|
; RV32IF-NEXT: addi a5, a3, -1
|
|
; RV32IF-NEXT: addi a3, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a5
|
|
; RV32IF-NEXT: add a0, a1, a4
|
|
; RV32IF-NEXT: seqz a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a5
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a0, a0, a1
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB20_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a1, 1043
|
|
; RV32IF-NEXT: li a4, 1075
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: sub a2, a4, a2
|
|
; RV32IF-NEXT: srai a4, a1, 31
|
|
; RV32IF-NEXT: srl a5, a0, a2
|
|
; RV32IF-NEXT: and s3, a4, a5
|
|
; RV32IF-NEXT: bltz a1, .LBB20_7
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl s4, a0, a1
|
|
; RV32IF-NEXT: j .LBB20_8
|
|
; RV32IF-NEXT: .LBB20_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: sw a3, 16(sp)
|
|
; RV32IF-NEXT: sw a0, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 0(a1)
|
|
; RV32IF-NEXT: lw a2, 4(a1)
|
|
; RV32IF-NEXT: lw a5, 8(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a6, a4, 1
|
|
; RV32IF-NEXT: srli a7, a2, 1
|
|
; RV32IF-NEXT: srli t0, a5, 1
|
|
; RV32IF-NEXT: srl a6, a6, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl a3, t0, a3
|
|
; RV32IF-NEXT: sll a2, a2, a0
|
|
; RV32IF-NEXT: sll a5, a5, a0
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or s3, a2, a6
|
|
; RV32IF-NEXT: or a2, a5, a7
|
|
; RV32IF-NEXT: or a3, a1, a3
|
|
; RV32IF-NEXT: sll s4, a4, a0
|
|
; RV32IF-NEXT: mv a0, s2
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s7, a0, s1
|
|
; RV32IF-NEXT: sltu a0, s7, a0
|
|
; RV32IF-NEXT: add s8, a1, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: add s2, s5, a1
|
|
; RV32IF-NEXT: sltu a0, s2, s5
|
|
; RV32IF-NEXT: add s5, s6, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s6, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s6, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s2, s5, a0
|
|
; RV32IF-NEXT: sltu s4, s2, s5
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s4
|
|
; RV32IF-NEXT: add s2, a0, s2
|
|
; RV32IF-NEXT: sltu a2, s2, a0
|
|
; RV32IF-NEXT: add a0, s2, s7
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: add a1, a1, s8
|
|
; RV32IF-NEXT: sltu a2, a0, s2
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: bnez a1, .LBB20_9
|
|
; RV32IF-NEXT: .LBB20_6:
|
|
; RV32IF-NEXT: seqz a2, a0
|
|
; RV32IF-NEXT: j .LBB20_10
|
|
; RV32IF-NEXT: .LBB20_7:
|
|
; RV32IF-NEXT: srl a1, a3, a2
|
|
; RV32IF-NEXT: not a2, a2
|
|
; RV32IF-NEXT: slli a0, a0, 1
|
|
; RV32IF-NEXT: sll a0, a0, a2
|
|
; RV32IF-NEXT: or s4, a1, a0
|
|
; RV32IF-NEXT: .LBB20_8: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s5, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s5, a0
|
|
; RV32IF-NEXT: add s7, a1, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s6, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s6, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s8, s7, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s2, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: add s9, a0, s8
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a0, s9, a0
|
|
; RV32IF-NEXT: sltu a2, s9, s2
|
|
; RV32IF-NEXT: sltu a3, s8, s7
|
|
; RV32IF-NEXT: sltu a4, a0, s9
|
|
; RV32IF-NEXT: add a3, s5, a3
|
|
; RV32IF-NEXT: add a2, a3, a2
|
|
; RV32IF-NEXT: add a1, a2, a1
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: beqz a1, .LBB20_6
|
|
; RV32IF-NEXT: .LBB20_9: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a2, a1, 31
|
|
; RV32IF-NEXT: .LBB20_10: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: xori a3, a0, 1
|
|
; RV32IF-NEXT: or a3, a3, a1
|
|
; RV32IF-NEXT: seqz a3, a3
|
|
; RV32IF-NEXT: addi a3, a3, -1
|
|
; RV32IF-NEXT: and a3, a3, a2
|
|
; RV32IF-NEXT: neg a2, a3
|
|
; RV32IF-NEXT: bnez a3, .LBB20_12
|
|
; RV32IF-NEXT: # %bb.11: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: li a0, 1
|
|
; RV32IF-NEXT: .LBB20_12: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: and a4, a2, a1
|
|
; RV32IF-NEXT: beqz a4, .LBB20_14
|
|
; RV32IF-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: sgtz a1, a4
|
|
; RV32IF-NEXT: j .LBB20_15
|
|
; RV32IF-NEXT: .LBB20_14:
|
|
; RV32IF-NEXT: snez a1, a0
|
|
; RV32IF-NEXT: .LBB20_15: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: and a3, a2, s6
|
|
; RV32IF-NEXT: or a0, a0, a4
|
|
; RV32IF-NEXT: and a2, a2, s1
|
|
; RV32IF-NEXT: bnez a0, .LBB20_17
|
|
; RV32IF-NEXT: # %bb.16:
|
|
; RV32IF-NEXT: or a0, a2, a3
|
|
; RV32IF-NEXT: snez a1, a0
|
|
; RV32IF-NEXT: .LBB20_17: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: neg a1, a1
|
|
; RV32IF-NEXT: and a0, a1, a2
|
|
; RV32IF-NEXT: and a1, a1, a3
|
|
; RV32IF-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s9, 36(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: .cfi_restore s0
|
|
; RV32IF-NEXT: .cfi_restore s1
|
|
; RV32IF-NEXT: .cfi_restore s2
|
|
; RV32IF-NEXT: .cfi_restore s3
|
|
; RV32IF-NEXT: .cfi_restore s4
|
|
; RV32IF-NEXT: .cfi_restore s5
|
|
; RV32IF-NEXT: .cfi_restore s6
|
|
; RV32IF-NEXT: .cfi_restore s7
|
|
; RV32IF-NEXT: .cfi_restore s8
|
|
; RV32IF-NEXT: .cfi_restore s9
|
|
; RV32IF-NEXT: addi sp, sp, 80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f64i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixdfti
|
|
; RV64-NEXT: slti a2, a1, 1
|
|
; RV64-NEXT: blez a1, .LBB20_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a1, 1
|
|
; RV64-NEXT: .LBB20_2: # %entry
|
|
; RV64-NEXT: neg a2, a2
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: beqz a1, .LBB20_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: sgtz a1, a1
|
|
; RV64-NEXT: j .LBB20_5
|
|
; RV64-NEXT: .LBB20_4:
|
|
; RV64-NEXT: snez a1, a0
|
|
; RV64-NEXT: .LBB20_5: # %entry
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i64:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 96
|
|
; RV32IFD-NEXT: sw ra, 92(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s0, 88(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s1, 84(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s2, 80(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s3, 76(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s4, 72(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s5, 68(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s6, 64(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s7, 60(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s8, 56(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s9, 52(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: .cfi_offset ra, -4
|
|
; RV32IFD-NEXT: .cfi_offset s0, -8
|
|
; RV32IFD-NEXT: .cfi_offset s1, -12
|
|
; RV32IFD-NEXT: .cfi_offset s2, -16
|
|
; RV32IFD-NEXT: .cfi_offset s3, -20
|
|
; RV32IFD-NEXT: .cfi_offset s4, -24
|
|
; RV32IFD-NEXT: .cfi_offset s5, -28
|
|
; RV32IFD-NEXT: .cfi_offset s6, -32
|
|
; RV32IFD-NEXT: .cfi_offset s7, -36
|
|
; RV32IFD-NEXT: .cfi_offset s8, -40
|
|
; RV32IFD-NEXT: .cfi_offset s9, -44
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB20_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li s1, 0
|
|
; RV32IFD-NEXT: li s6, 0
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: j .LBB20_6
|
|
; RV32IFD-NEXT: .LBB20_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: srai s0, a1, 31
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: ori s2, s0, 1
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB20_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 1043
|
|
; RV32IFD-NEXT: li a4, 1075
|
|
; RV32IFD-NEXT: sub a3, a3, a0
|
|
; RV32IFD-NEXT: sub a0, a4, a0
|
|
; RV32IFD-NEXT: srai a4, a3, 31
|
|
; RV32IFD-NEXT: srl a5, a1, a0
|
|
; RV32IFD-NEXT: and s3, a4, a5
|
|
; RV32IFD-NEXT: bltz a3, .LBB20_7
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl s4, a1, a3
|
|
; RV32IFD-NEXT: j .LBB20_8
|
|
; RV32IFD-NEXT: .LBB20_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 0(a1)
|
|
; RV32IFD-NEXT: lw a2, 4(a1)
|
|
; RV32IFD-NEXT: lw a5, 8(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a6, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a2, 1
|
|
; RV32IFD-NEXT: srli t0, a5, 1
|
|
; RV32IFD-NEXT: srl a6, a6, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl a3, t0, a3
|
|
; RV32IFD-NEXT: sll a2, a2, a0
|
|
; RV32IFD-NEXT: sll a5, a5, a0
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or s3, a2, a6
|
|
; RV32IFD-NEXT: or a2, a5, a7
|
|
; RV32IFD-NEXT: or a3, a1, a3
|
|
; RV32IFD-NEXT: sll s4, a4, a0
|
|
; RV32IFD-NEXT: mv a0, s2
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s7, a0, s1
|
|
; RV32IFD-NEXT: sltu a0, s7, a0
|
|
; RV32IFD-NEXT: add s8, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: add s2, s5, a1
|
|
; RV32IFD-NEXT: sltu a0, s2, s5
|
|
; RV32IFD-NEXT: add s5, s6, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s6, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s6, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s2, s5, a0
|
|
; RV32IFD-NEXT: sltu s4, s2, s5
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s4
|
|
; RV32IFD-NEXT: add s2, a0, s2
|
|
; RV32IFD-NEXT: sltu a2, s2, a0
|
|
; RV32IFD-NEXT: add a0, s2, s7
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: add a1, a1, s8
|
|
; RV32IFD-NEXT: sltu a2, a0, s2
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: bnez a1, .LBB20_9
|
|
; RV32IFD-NEXT: .LBB20_6:
|
|
; RV32IFD-NEXT: seqz a2, a0
|
|
; RV32IFD-NEXT: j .LBB20_10
|
|
; RV32IFD-NEXT: .LBB20_7:
|
|
; RV32IFD-NEXT: srl a2, a2, a0
|
|
; RV32IFD-NEXT: not a0, a0
|
|
; RV32IFD-NEXT: slli a1, a1, 1
|
|
; RV32IFD-NEXT: sll a0, a1, a0
|
|
; RV32IFD-NEXT: or s4, a2, a0
|
|
; RV32IFD-NEXT: .LBB20_8: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s5, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s5, a0
|
|
; RV32IFD-NEXT: add s7, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s6, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s6, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s8, s7, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s2, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: add s9, a0, s8
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a0, s9, a0
|
|
; RV32IFD-NEXT: sltu a2, s9, s2
|
|
; RV32IFD-NEXT: sltu a3, s8, s7
|
|
; RV32IFD-NEXT: sltu a4, a0, s9
|
|
; RV32IFD-NEXT: add a3, s5, a3
|
|
; RV32IFD-NEXT: add a2, a3, a2
|
|
; RV32IFD-NEXT: add a1, a2, a1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: beqz a1, .LBB20_6
|
|
; RV32IFD-NEXT: .LBB20_9: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a2, a1, 31
|
|
; RV32IFD-NEXT: .LBB20_10: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: xori a3, a0, 1
|
|
; RV32IFD-NEXT: or a3, a3, a1
|
|
; RV32IFD-NEXT: seqz a3, a3
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a3, a3, a2
|
|
; RV32IFD-NEXT: neg a2, a3
|
|
; RV32IFD-NEXT: bnez a3, .LBB20_12
|
|
; RV32IFD-NEXT: # %bb.11: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: li a0, 1
|
|
; RV32IFD-NEXT: .LBB20_12: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: and a4, a2, a1
|
|
; RV32IFD-NEXT: beqz a4, .LBB20_14
|
|
; RV32IFD-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: sgtz a1, a4
|
|
; RV32IFD-NEXT: j .LBB20_15
|
|
; RV32IFD-NEXT: .LBB20_14:
|
|
; RV32IFD-NEXT: snez a1, a0
|
|
; RV32IFD-NEXT: .LBB20_15: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: and a3, a2, s6
|
|
; RV32IFD-NEXT: or a0, a0, a4
|
|
; RV32IFD-NEXT: and a2, a2, s1
|
|
; RV32IFD-NEXT: bnez a0, .LBB20_17
|
|
; RV32IFD-NEXT: # %bb.16:
|
|
; RV32IFD-NEXT: or a0, a2, a3
|
|
; RV32IFD-NEXT: snez a1, a0
|
|
; RV32IFD-NEXT: .LBB20_17: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: neg a1, a1
|
|
; RV32IFD-NEXT: and a0, a1, a2
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: lw ra, 92(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s0, 88(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s1, 84(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s2, 80(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s3, 76(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s4, 72(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s5, 68(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s6, 64(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s7, 60(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s8, 56(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s9, 52(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: .cfi_restore ra
|
|
; RV32IFD-NEXT: .cfi_restore s0
|
|
; RV32IFD-NEXT: .cfi_restore s1
|
|
; RV32IFD-NEXT: .cfi_restore s2
|
|
; RV32IFD-NEXT: .cfi_restore s3
|
|
; RV32IFD-NEXT: .cfi_restore s4
|
|
; RV32IFD-NEXT: .cfi_restore s5
|
|
; RV32IFD-NEXT: .cfi_restore s6
|
|
; RV32IFD-NEXT: .cfi_restore s7
|
|
; RV32IFD-NEXT: .cfi_restore s8
|
|
; RV32IFD-NEXT: .cfi_restore s9
|
|
; RV32IFD-NEXT: addi sp, sp, 96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i128
|
|
%0 = icmp slt i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%1 = icmp sgt i128 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 0
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @stest_f32i64(float %x) {
|
|
; RV32-LABEL: stest_f32i64:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: addi sp, sp, -80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: .cfi_offset s0, -8
|
|
; RV32-NEXT: .cfi_offset s1, -12
|
|
; RV32-NEXT: .cfi_offset s2, -16
|
|
; RV32-NEXT: .cfi_offset s3, -20
|
|
; RV32-NEXT: .cfi_offset s4, -24
|
|
; RV32-NEXT: .cfi_offset s5, -28
|
|
; RV32-NEXT: .cfi_offset s6, -32
|
|
; RV32-NEXT: .cfi_offset s7, -36
|
|
; RV32-NEXT: .cfi_offset s8, -40
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB21_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li s1, 0
|
|
; RV32-NEXT: li s2, 0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: j .LBB21_5
|
|
; RV32-NEXT: .LBB21_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: srai s0, a1, 31
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: or a1, a1, a2
|
|
; RV32-NEXT: ori s2, s0, 1
|
|
; RV32-NEXT: bltu a3, a0, .LBB21_4
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a2, 150
|
|
; RV32-NEXT: sub a2, a2, a0
|
|
; RV32-NEXT: srl s3, a1, a2
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s4
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add s6, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s7, s6, a0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s4, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: add s8, a0, s7
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s3
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a0, s8, a0
|
|
; RV32-NEXT: sltu a2, s8, s4
|
|
; RV32-NEXT: sltu a3, s7, s6
|
|
; RV32-NEXT: sltu a4, a0, s8
|
|
; RV32-NEXT: add a3, s5, a3
|
|
; RV32-NEXT: add a2, a3, a2
|
|
; RV32-NEXT: add a1, a2, a1
|
|
; RV32-NEXT: add a1, a1, a4
|
|
; RV32-NEXT: j .LBB21_5
|
|
; RV32-NEXT: .LBB21_4: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a1, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 32(sp)
|
|
; RV32-NEXT: sw zero, 36(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: sw zero, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: addi a1, sp, 24
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 0(a1)
|
|
; RV32-NEXT: lw a2, 4(a1)
|
|
; RV32-NEXT: lw a5, 8(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a6, a4, 1
|
|
; RV32-NEXT: srli a7, a2, 1
|
|
; RV32-NEXT: srli t0, a5, 1
|
|
; RV32-NEXT: srl a6, a6, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl a3, t0, a3
|
|
; RV32-NEXT: sll a2, a2, a0
|
|
; RV32-NEXT: sll a5, a5, a0
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or s3, a2, a6
|
|
; RV32-NEXT: or a2, a5, a7
|
|
; RV32-NEXT: or a3, a1, a3
|
|
; RV32-NEXT: sll s4, a4, a0
|
|
; RV32-NEXT: mv a0, s2
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s4
|
|
; RV32-NEXT: mv a3, s3
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s7, a0, s1
|
|
; RV32-NEXT: sltu a0, s7, a0
|
|
; RV32-NEXT: add s8, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s5, a0
|
|
; RV32-NEXT: mv s6, a1
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: add s2, s5, a1
|
|
; RV32-NEXT: sltu a0, s2, s5
|
|
; RV32-NEXT: add s6, s6, a0
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s4, s6, a0
|
|
; RV32-NEXT: sltu s5, s4, s6
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s4, a0, s4
|
|
; RV32-NEXT: sltu a2, s4, a0
|
|
; RV32-NEXT: add a0, s4, s7
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: add a1, a1, s8
|
|
; RV32-NEXT: sltu a2, a0, s4
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: .LBB21_5: # %fp-to-i-cleanup
|
|
; RV32-NEXT: lui a2, 524288
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: beq s2, a3, .LBB21_7
|
|
; RV32-NEXT: # %bb.6: # %fp-to-i-cleanup
|
|
; RV32-NEXT: sltu a4, s2, a3
|
|
; RV32-NEXT: or a5, a0, a1
|
|
; RV32-NEXT: bnez a5, .LBB21_8
|
|
; RV32-NEXT: j .LBB21_9
|
|
; RV32-NEXT: .LBB21_7:
|
|
; RV32-NEXT: sltiu a4, s1, -1
|
|
; RV32-NEXT: or a5, a0, a1
|
|
; RV32-NEXT: beqz a5, .LBB21_9
|
|
; RV32-NEXT: .LBB21_8: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a4, a1, 31
|
|
; RV32-NEXT: .LBB21_9: # %fp-to-i-cleanup
|
|
; RV32-NEXT: neg a5, a4
|
|
; RV32-NEXT: addi a6, a4, -1
|
|
; RV32-NEXT: bnez a4, .LBB21_11
|
|
; RV32-NEXT: # %bb.10: # %fp-to-i-cleanup
|
|
; RV32-NEXT: mv s2, a3
|
|
; RV32-NEXT: .LBB21_11: # %fp-to-i-cleanup
|
|
; RV32-NEXT: or a3, a6, s1
|
|
; RV32-NEXT: and a1, a5, a1
|
|
; RV32-NEXT: and a4, a5, a0
|
|
; RV32-NEXT: beq s2, a2, .LBB21_13
|
|
; RV32-NEXT: # %bb.12: # %fp-to-i-cleanup
|
|
; RV32-NEXT: sltu a0, a2, s2
|
|
; RV32-NEXT: j .LBB21_14
|
|
; RV32-NEXT: .LBB21_13:
|
|
; RV32-NEXT: snez a0, a3
|
|
; RV32-NEXT: .LBB21_14: # %fp-to-i-cleanup
|
|
; RV32-NEXT: and a4, a4, a1
|
|
; RV32-NEXT: li a2, -1
|
|
; RV32-NEXT: beq a4, a2, .LBB21_16
|
|
; RV32-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a1, a1, 31
|
|
; RV32-NEXT: xori a0, a1, 1
|
|
; RV32-NEXT: .LBB21_16: # %fp-to-i-cleanup
|
|
; RV32-NEXT: bnez a0, .LBB21_18
|
|
; RV32-NEXT: # %bb.17: # %fp-to-i-cleanup
|
|
; RV32-NEXT: lui s2, 524288
|
|
; RV32-NEXT: .LBB21_18: # %fp-to-i-cleanup
|
|
; RV32-NEXT: neg a0, a0
|
|
; RV32-NEXT: and a0, a0, a3
|
|
; RV32-NEXT: mv a1, s2
|
|
; RV32-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: .cfi_restore s0
|
|
; RV32-NEXT: .cfi_restore s1
|
|
; RV32-NEXT: .cfi_restore s2
|
|
; RV32-NEXT: .cfi_restore s3
|
|
; RV32-NEXT: .cfi_restore s4
|
|
; RV32-NEXT: .cfi_restore s5
|
|
; RV32-NEXT: .cfi_restore s6
|
|
; RV32-NEXT: .cfi_restore s7
|
|
; RV32-NEXT: .cfi_restore s8
|
|
; RV32-NEXT: addi sp, sp, 80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: feq.s a1, fa0, fa0
|
|
; RV64-NEXT: seqz a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i128
|
|
%0 = icmp slt i128 %conv, 9223372036854775807
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 9223372036854775807
|
|
%1 = icmp sgt i128 %spec.store.select, -9223372036854775808
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 -9223372036854775808
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utest_f32i64(float %x) {
|
|
; RV32-LABEL: utest_f32i64:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB22_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: li a2, 0
|
|
; RV32-NEXT: j .LBB22_5
|
|
; RV32-NEXT: .LBB22_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: or a4, a1, a2
|
|
; RV32-NEXT: bltu a3, a0, .LBB22_4
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: li a2, 0
|
|
; RV32-NEXT: li a5, 150
|
|
; RV32-NEXT: sub a5, a5, a0
|
|
; RV32-NEXT: srl a0, a4, a5
|
|
; RV32-NEXT: j .LBB22_5
|
|
; RV32-NEXT: .LBB22_4: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi sp, sp, -32
|
|
; RV32-NEXT: .cfi_def_cfa_offset 32
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a4, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: sw zero, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 0(sp)
|
|
; RV32-NEXT: sw zero, 4(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: addi a1, sp, 16
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 8(a1)
|
|
; RV32-NEXT: lw a5, 4(a1)
|
|
; RV32-NEXT: lw a6, 0(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a2, a4, 1
|
|
; RV32-NEXT: srli a7, a5, 1
|
|
; RV32-NEXT: srli t0, a6, 1
|
|
; RV32-NEXT: srl a2, a2, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl t0, t0, a3
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or a2, a1, a2
|
|
; RV32-NEXT: sll a1, a4, a0
|
|
; RV32-NEXT: sll a4, a5, a0
|
|
; RV32-NEXT: or a3, a1, a7
|
|
; RV32-NEXT: or a1, a4, t0
|
|
; RV32-NEXT: sll a0, a6, a0
|
|
; RV32-NEXT: addi sp, sp, 32
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: .LBB22_5: # %fp-to-i-cleanup
|
|
; RV32-NEXT: or a4, a2, a3
|
|
; RV32-NEXT: xori a3, a3, 1
|
|
; RV32-NEXT: seqz a4, a4
|
|
; RV32-NEXT: or a2, a3, a2
|
|
; RV32-NEXT: seqz a2, a2
|
|
; RV32-NEXT: addi a2, a2, -1
|
|
; RV32-NEXT: and a2, a2, a4
|
|
; RV32-NEXT: neg a2, a2
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixunssfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i128
|
|
%0 = icmp ult i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f32i64(float %x) {
|
|
; RV32-LABEL: ustest_f32i64:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: addi sp, sp, -80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: .cfi_offset s0, -8
|
|
; RV32-NEXT: .cfi_offset s1, -12
|
|
; RV32-NEXT: .cfi_offset s2, -16
|
|
; RV32-NEXT: .cfi_offset s3, -20
|
|
; RV32-NEXT: .cfi_offset s4, -24
|
|
; RV32-NEXT: .cfi_offset s5, -28
|
|
; RV32-NEXT: .cfi_offset s6, -32
|
|
; RV32-NEXT: .cfi_offset s7, -36
|
|
; RV32-NEXT: .cfi_offset s8, -40
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB23_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li s1, 0
|
|
; RV32-NEXT: li s5, 0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: j .LBB23_4
|
|
; RV32-NEXT: .LBB23_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: srai s0, a1, 31
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: or a1, a1, a2
|
|
; RV32-NEXT: ori s2, s0, 1
|
|
; RV32-NEXT: bltu a3, a0, .LBB23_5
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a2, 150
|
|
; RV32-NEXT: sub a2, a2, a0
|
|
; RV32-NEXT: srl s3, a1, a2
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s4, a0, s4
|
|
; RV32-NEXT: sltu a0, s4, a0
|
|
; RV32-NEXT: add s6, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s5, a0, s4
|
|
; RV32-NEXT: sltu a0, s5, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s7, s6, a0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s2, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: add s8, a0, s7
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s3
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a0, s8, a0
|
|
; RV32-NEXT: sltu a2, s8, s2
|
|
; RV32-NEXT: sltu a3, s7, s6
|
|
; RV32-NEXT: sltu a4, a0, s8
|
|
; RV32-NEXT: add a3, s4, a3
|
|
; RV32-NEXT: add a2, a3, a2
|
|
; RV32-NEXT: add a1, a2, a1
|
|
; RV32-NEXT: add a1, a1, a4
|
|
; RV32-NEXT: bnez a1, .LBB23_6
|
|
; RV32-NEXT: .LBB23_4:
|
|
; RV32-NEXT: seqz a2, a0
|
|
; RV32-NEXT: j .LBB23_7
|
|
; RV32-NEXT: .LBB23_5: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a1, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 32(sp)
|
|
; RV32-NEXT: sw zero, 36(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: sw zero, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: addi a1, sp, 24
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 0(a1)
|
|
; RV32-NEXT: lw a2, 4(a1)
|
|
; RV32-NEXT: lw a5, 8(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a6, a4, 1
|
|
; RV32-NEXT: srli a7, a2, 1
|
|
; RV32-NEXT: srli t0, a5, 1
|
|
; RV32-NEXT: srl a6, a6, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl a3, t0, a3
|
|
; RV32-NEXT: sll a2, a2, a0
|
|
; RV32-NEXT: sll a5, a5, a0
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or s3, a2, a6
|
|
; RV32-NEXT: or a2, a5, a7
|
|
; RV32-NEXT: or a3, a1, a3
|
|
; RV32-NEXT: sll s4, a4, a0
|
|
; RV32-NEXT: mv a0, s2
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s4
|
|
; RV32-NEXT: mv a3, s3
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s7, a0, s1
|
|
; RV32-NEXT: sltu a0, s7, a0
|
|
; RV32-NEXT: add s8, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s5, a0
|
|
; RV32-NEXT: mv s6, a1
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: add s2, s5, a1
|
|
; RV32-NEXT: sltu a0, s2, s5
|
|
; RV32-NEXT: add s6, s6, a0
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s5, a0, s2
|
|
; RV32-NEXT: sltu a0, s5, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s2, s6, a0
|
|
; RV32-NEXT: sltu s4, s2, s6
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s4
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a2, s2, a0
|
|
; RV32-NEXT: add a0, s2, s7
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: add a1, a1, s8
|
|
; RV32-NEXT: sltu a2, a0, s2
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: beqz a1, .LBB23_4
|
|
; RV32-NEXT: .LBB23_6: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: .LBB23_7: # %fp-to-i-cleanup
|
|
; RV32-NEXT: xori a3, a0, 1
|
|
; RV32-NEXT: or a3, a3, a1
|
|
; RV32-NEXT: seqz a3, a3
|
|
; RV32-NEXT: addi a3, a3, -1
|
|
; RV32-NEXT: and a3, a3, a2
|
|
; RV32-NEXT: neg a2, a3
|
|
; RV32-NEXT: bnez a3, .LBB23_9
|
|
; RV32-NEXT: # %bb.8: # %fp-to-i-cleanup
|
|
; RV32-NEXT: li a0, 1
|
|
; RV32-NEXT: .LBB23_9: # %fp-to-i-cleanup
|
|
; RV32-NEXT: and a4, a2, a1
|
|
; RV32-NEXT: beqz a4, .LBB23_11
|
|
; RV32-NEXT: # %bb.10: # %fp-to-i-cleanup
|
|
; RV32-NEXT: sgtz a1, a4
|
|
; RV32-NEXT: j .LBB23_12
|
|
; RV32-NEXT: .LBB23_11:
|
|
; RV32-NEXT: snez a1, a0
|
|
; RV32-NEXT: .LBB23_12: # %fp-to-i-cleanup
|
|
; RV32-NEXT: and a3, a2, s5
|
|
; RV32-NEXT: or a0, a0, a4
|
|
; RV32-NEXT: and a2, a2, s1
|
|
; RV32-NEXT: bnez a0, .LBB23_14
|
|
; RV32-NEXT: # %bb.13:
|
|
; RV32-NEXT: or a0, a2, a3
|
|
; RV32-NEXT: snez a1, a0
|
|
; RV32-NEXT: .LBB23_14: # %fp-to-i-cleanup
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a2
|
|
; RV32-NEXT: and a1, a1, a3
|
|
; RV32-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: .cfi_restore s0
|
|
; RV32-NEXT: .cfi_restore s1
|
|
; RV32-NEXT: .cfi_restore s2
|
|
; RV32-NEXT: .cfi_restore s3
|
|
; RV32-NEXT: .cfi_restore s4
|
|
; RV32-NEXT: .cfi_restore s5
|
|
; RV32-NEXT: .cfi_restore s6
|
|
; RV32-NEXT: .cfi_restore s7
|
|
; RV32-NEXT: .cfi_restore s8
|
|
; RV32-NEXT: addi sp, sp, 80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: slti a2, a1, 1
|
|
; RV64-NEXT: blez a1, .LBB23_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a1, 1
|
|
; RV64-NEXT: .LBB23_2: # %entry
|
|
; RV64-NEXT: neg a2, a2
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: beqz a1, .LBB23_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: sgtz a1, a1
|
|
; RV64-NEXT: j .LBB23_5
|
|
; RV64-NEXT: .LBB23_4:
|
|
; RV64-NEXT: snez a1, a0
|
|
; RV64-NEXT: .LBB23_5: # %entry
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i128
|
|
%0 = icmp slt i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%1 = icmp sgt i128 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 0
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @stest_f16i64(half %x) {
|
|
; RV32-LABEL: stest_f16i64:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a2, fa0, rtz
|
|
; RV32-NEXT: lui a0, 524288
|
|
; RV32-NEXT: srai a3, a2, 31
|
|
; RV32-NEXT: addi a4, a0, -1
|
|
; RV32-NEXT: beq a3, a4, .LBB24_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: sltu a5, a3, a4
|
|
; RV32-NEXT: bnez a3, .LBB24_3
|
|
; RV32-NEXT: j .LBB24_4
|
|
; RV32-NEXT: .LBB24_2:
|
|
; RV32-NEXT: sltiu a5, a2, -1
|
|
; RV32-NEXT: beqz a3, .LBB24_4
|
|
; RV32-NEXT: .LBB24_3: # %entry
|
|
; RV32-NEXT: srli a5, a3, 31
|
|
; RV32-NEXT: .LBB24_4: # %entry
|
|
; RV32-NEXT: neg a6, a5
|
|
; RV32-NEXT: addi a7, a5, -1
|
|
; RV32-NEXT: mv a1, a3
|
|
; RV32-NEXT: bnez a5, .LBB24_6
|
|
; RV32-NEXT: # %bb.5: # %entry
|
|
; RV32-NEXT: mv a1, a4
|
|
; RV32-NEXT: .LBB24_6: # %entry
|
|
; RV32-NEXT: or a2, a7, a2
|
|
; RV32-NEXT: and a3, a6, a3
|
|
; RV32-NEXT: beq a1, a0, .LBB24_8
|
|
; RV32-NEXT: # %bb.7: # %entry
|
|
; RV32-NEXT: sltu a0, a0, a1
|
|
; RV32-NEXT: li a4, -1
|
|
; RV32-NEXT: bne a3, a4, .LBB24_9
|
|
; RV32-NEXT: j .LBB24_10
|
|
; RV32-NEXT: .LBB24_8:
|
|
; RV32-NEXT: snez a0, a2
|
|
; RV32-NEXT: li a4, -1
|
|
; RV32-NEXT: beq a3, a4, .LBB24_10
|
|
; RV32-NEXT: .LBB24_9: # %entry
|
|
; RV32-NEXT: srli a3, a3, 31
|
|
; RV32-NEXT: xori a0, a3, 1
|
|
; RV32-NEXT: .LBB24_10: # %entry
|
|
; RV32-NEXT: bnez a0, .LBB24_12
|
|
; RV32-NEXT: # %bb.11: # %entry
|
|
; RV32-NEXT: lui a1, 524288
|
|
; RV32-NEXT: .LBB24_12: # %entry
|
|
; RV32-NEXT: neg a0, a0
|
|
; RV32-NEXT: and a0, a0, a2
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: li a2, -1
|
|
; RV64-NEXT: srli a3, a2, 1
|
|
; RV64-NEXT: beqz a1, .LBB24_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: srli a4, a1, 63
|
|
; RV64-NEXT: j .LBB24_3
|
|
; RV64-NEXT: .LBB24_2:
|
|
; RV64-NEXT: sltu a4, a0, a3
|
|
; RV64-NEXT: .LBB24_3: # %entry
|
|
; RV64-NEXT: neg a5, a4
|
|
; RV64-NEXT: and a5, a5, a1
|
|
; RV64-NEXT: bnez a4, .LBB24_5
|
|
; RV64-NEXT: # %bb.4: # %entry
|
|
; RV64-NEXT: mv a0, a3
|
|
; RV64-NEXT: .LBB24_5: # %entry
|
|
; RV64-NEXT: slli a1, a2, 63
|
|
; RV64-NEXT: beq a5, a2, .LBB24_7
|
|
; RV64-NEXT: # %bb.6: # %entry
|
|
; RV64-NEXT: srli a5, a5, 63
|
|
; RV64-NEXT: xori a2, a5, 1
|
|
; RV64-NEXT: beqz a2, .LBB24_8
|
|
; RV64-NEXT: j .LBB24_9
|
|
; RV64-NEXT: .LBB24_7:
|
|
; RV64-NEXT: sltu a2, a1, a0
|
|
; RV64-NEXT: bnez a2, .LBB24_9
|
|
; RV64-NEXT: .LBB24_8: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB24_9: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i128
|
|
%0 = icmp slt i128 %conv, 9223372036854775807
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 9223372036854775807
|
|
%1 = icmp sgt i128 %spec.store.select, -9223372036854775808
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 -9223372036854775808
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utesth_f16i64(half %x) {
|
|
; RV32-LABEL: utesth_f16i64:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utesth_f16i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixunssfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i128
|
|
%0 = icmp ult i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f16i64(half %x) {
|
|
; RV32-LABEL: ustest_f16i64:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: srai a1, a0, 31
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: seqz a3, a1
|
|
; RV32-NEXT: ori a4, a1, 1
|
|
; RV32-NEXT: or a2, a3, a2
|
|
; RV32-NEXT: seqz a3, a4
|
|
; RV32-NEXT: addi a3, a3, -1
|
|
; RV32-NEXT: and a4, a3, a2
|
|
; RV32-NEXT: neg a2, a4
|
|
; RV32-NEXT: mv a3, a1
|
|
; RV32-NEXT: bnez a4, .LBB26_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: li a3, 1
|
|
; RV32-NEXT: .LBB26_2: # %entry
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: beqz a1, .LBB26_4
|
|
; RV32-NEXT: # %bb.3: # %entry
|
|
; RV32-NEXT: sgtz a4, a1
|
|
; RV32-NEXT: j .LBB26_5
|
|
; RV32-NEXT: .LBB26_4:
|
|
; RV32-NEXT: snez a4, a3
|
|
; RV32-NEXT: .LBB26_5: # %entry
|
|
; RV32-NEXT: or a3, a3, a1
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: bnez a3, .LBB26_7
|
|
; RV32-NEXT: # %bb.6:
|
|
; RV32-NEXT: or a2, a0, a1
|
|
; RV32-NEXT: snez a4, a2
|
|
; RV32-NEXT: .LBB26_7: # %entry
|
|
; RV32-NEXT: neg a2, a4
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i64:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: slti a2, a1, 1
|
|
; RV64-NEXT: blez a1, .LBB26_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a1, 1
|
|
; RV64-NEXT: .LBB26_2: # %entry
|
|
; RV64-NEXT: neg a2, a2
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: beqz a1, .LBB26_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: sgtz a1, a1
|
|
; RV64-NEXT: j .LBB26_5
|
|
; RV64-NEXT: .LBB26_4:
|
|
; RV64-NEXT: snez a1, a0
|
|
; RV64-NEXT: .LBB26_5: # %entry
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i128
|
|
%0 = icmp slt i128 %conv, 18446744073709551616
|
|
%spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
|
|
%1 = icmp sgt i128 %spec.store.select, 0
|
|
%spec.store.select7 = select i1 %1, i128 %spec.store.select, i128 0
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
; i32 saturate
|
|
|
|
define i32 @stest_f64i32_mm(double %x) {
|
|
; RV32IF-LABEL: stest_f64i32_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfdi
|
|
; RV32IF-NEXT: lui a2, 524288
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: beqz a1, .LBB27_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: srli a4, a1, 31
|
|
; RV32IF-NEXT: j .LBB27_3
|
|
; RV32IF-NEXT: .LBB27_2:
|
|
; RV32IF-NEXT: sltu a4, a0, a3
|
|
; RV32IF-NEXT: .LBB27_3: # %entry
|
|
; RV32IF-NEXT: neg a5, a4
|
|
; RV32IF-NEXT: and a1, a5, a1
|
|
; RV32IF-NEXT: bnez a4, .LBB27_5
|
|
; RV32IF-NEXT: # %bb.4: # %entry
|
|
; RV32IF-NEXT: mv a0, a3
|
|
; RV32IF-NEXT: .LBB27_5: # %entry
|
|
; RV32IF-NEXT: li a3, -1
|
|
; RV32IF-NEXT: beq a1, a3, .LBB27_7
|
|
; RV32IF-NEXT: # %bb.6: # %entry
|
|
; RV32IF-NEXT: srli a1, a1, 31
|
|
; RV32IF-NEXT: xori a1, a1, 1
|
|
; RV32IF-NEXT: beqz a1, .LBB27_8
|
|
; RV32IF-NEXT: j .LBB27_9
|
|
; RV32IF-NEXT: .LBB27_7:
|
|
; RV32IF-NEXT: sltu a1, a2, a0
|
|
; RV32IF-NEXT: bnez a1, .LBB27_9
|
|
; RV32IF-NEXT: .LBB27_8: # %entry
|
|
; RV32IF-NEXT: lui a0, 524288
|
|
; RV32IF-NEXT: .LBB27_9: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i32_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfdi
|
|
; RV64IF-NEXT: lui a1, 524288
|
|
; RV64IF-NEXT: addiw a2, a1, -1
|
|
; RV64IF-NEXT: blt a0, a2, .LBB27_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a2
|
|
; RV64IF-NEXT: .LBB27_2: # %entry
|
|
; RV64IF-NEXT: blt a1, a0, .LBB27_4
|
|
; RV64IF-NEXT: # %bb.3: # %entry
|
|
; RV64IF-NEXT: lui a0, 524288
|
|
; RV64IF-NEXT: .LBB27_4: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i32_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i32_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 524288
|
|
; RV64IFD-NEXT: addiw a2, a1, -1
|
|
; RV64IFD-NEXT: bge a0, a2, .LBB27_3
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: bge a1, a0, .LBB27_4
|
|
; RV64IFD-NEXT: .LBB27_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
; RV64IFD-NEXT: .LBB27_3: # %entry
|
|
; RV64IFD-NEXT: mv a0, a2
|
|
; RV64IFD-NEXT: blt a1, a2, .LBB27_2
|
|
; RV64IFD-NEXT: .LBB27_4: # %entry
|
|
; RV64IFD-NEXT: lui a0, 524288
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 2147483647)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 -2147483648)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utest_f64i32_mm(double %x) {
|
|
; RV32IF-LABEL: utest_f64i32_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixunsdfdi
|
|
; RV32IF-NEXT: seqz a1, a1
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: or a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: utest_f64i32_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixunsdfdi
|
|
; RV64IF-NEXT: li a1, -1
|
|
; RV64IF-NEXT: srli a1, a1, 32
|
|
; RV64IF-NEXT: bltu a0, a1, .LBB28_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB28_2: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i32_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: utest_f64i32_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.lu.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: li a1, -1
|
|
; RV64IFD-NEXT: srli a1, a1, 32
|
|
; RV64IFD-NEXT: bltu a0, a1, .LBB28_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB28_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i64
|
|
%spec.store.select = call i64 @llvm.umin.i64(i64 %conv, i64 4294967295)
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f64i32_mm(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i32_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfdi
|
|
; RV32IF-NEXT: bnez a1, .LBB29_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: li a2, 1
|
|
; RV32IF-NEXT: j .LBB29_3
|
|
; RV32IF-NEXT: .LBB29_2:
|
|
; RV32IF-NEXT: slti a2, a1, 1
|
|
; RV32IF-NEXT: .LBB29_3: # %entry
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: neg a2, a2
|
|
; RV32IF-NEXT: or a0, a3, a0
|
|
; RV32IF-NEXT: and a1, a2, a1
|
|
; RV32IF-NEXT: srli a1, a1, 31
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: and a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: ustest_f64i32_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfdi
|
|
; RV64IF-NEXT: li a1, -1
|
|
; RV64IF-NEXT: srli a1, a1, 32
|
|
; RV64IF-NEXT: blt a0, a1, .LBB29_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB29_2: # %entry
|
|
; RV64IF-NEXT: sgtz a1, a0
|
|
; RV64IF-NEXT: neg a1, a1
|
|
; RV64IF-NEXT: and a0, a1, a0
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i32_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV32IFD-NEXT: seqz a1, a1
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: ustest_f64i32_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: li a1, -1
|
|
; RV64IFD-NEXT: srli a1, a1, 32
|
|
; RV64IFD-NEXT: blt a0, a1, .LBB29_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB29_2: # %entry
|
|
; RV64IFD-NEXT: sgtz a1, a0
|
|
; RV64IFD-NEXT: neg a1, a1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 4294967295)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 0)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @stest_f32i32_mm(float %x) {
|
|
; RV32-LABEL: stest_f32i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 524288
|
|
; RV64-NEXT: addiw a2, a1, -1
|
|
; RV64-NEXT: bge a0, a2, .LBB30_3
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: bge a1, a0, .LBB30_4
|
|
; RV64-NEXT: .LBB30_2: # %entry
|
|
; RV64-NEXT: ret
|
|
; RV64-NEXT: .LBB30_3: # %entry
|
|
; RV64-NEXT: mv a0, a2
|
|
; RV64-NEXT: blt a1, a2, .LBB30_2
|
|
; RV64-NEXT: .LBB30_4: # %entry
|
|
; RV64-NEXT: lui a0, 524288
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 2147483647)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 -2147483648)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utest_f32i32_mm(float %x) {
|
|
; RV32-LABEL: utest_f32i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: bltu a0, a1, .LBB31_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB31_2: # %entry
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i64
|
|
%spec.store.select = call i64 @llvm.umin.i64(i64 %conv, i64 4294967295)
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f32i32_mm(float %x) {
|
|
; RV32-LABEL: ustest_f32i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: feq.s a1, fa0, fa0
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: blt a0, a1, .LBB32_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB32_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 4294967295)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 0)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @stest_f16i32_mm(half %x) {
|
|
; RV32-LABEL: stest_f16i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixsfdi
|
|
; RV32-NEXT: lui a2, 524288
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: beqz a1, .LBB33_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: srli a4, a1, 31
|
|
; RV32-NEXT: j .LBB33_3
|
|
; RV32-NEXT: .LBB33_2:
|
|
; RV32-NEXT: sltu a4, a0, a3
|
|
; RV32-NEXT: .LBB33_3: # %entry
|
|
; RV32-NEXT: neg a5, a4
|
|
; RV32-NEXT: and a1, a5, a1
|
|
; RV32-NEXT: bnez a4, .LBB33_5
|
|
; RV32-NEXT: # %bb.4: # %entry
|
|
; RV32-NEXT: mv a0, a3
|
|
; RV32-NEXT: .LBB33_5: # %entry
|
|
; RV32-NEXT: li a3, -1
|
|
; RV32-NEXT: beq a1, a3, .LBB33_7
|
|
; RV32-NEXT: # %bb.6: # %entry
|
|
; RV32-NEXT: srli a1, a1, 31
|
|
; RV32-NEXT: xori a1, a1, 1
|
|
; RV32-NEXT: beqz a1, .LBB33_8
|
|
; RV32-NEXT: j .LBB33_9
|
|
; RV32-NEXT: .LBB33_7:
|
|
; RV32-NEXT: sltu a1, a2, a0
|
|
; RV32-NEXT: bnez a1, .LBB33_9
|
|
; RV32-NEXT: .LBB33_8: # %entry
|
|
; RV32-NEXT: lui a0, 524288
|
|
; RV32-NEXT: .LBB33_9: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 524288
|
|
; RV64-NEXT: addiw a2, a1, -1
|
|
; RV64-NEXT: blt a0, a2, .LBB33_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a2
|
|
; RV64-NEXT: .LBB33_2: # %entry
|
|
; RV64-NEXT: blt a1, a0, .LBB33_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: lui a0, 524288
|
|
; RV64-NEXT: .LBB33_4: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 2147483647)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 -2147483648)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @utesth_f16i32_mm(half %x) {
|
|
; RV32-LABEL: utesth_f16i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixunssfdi
|
|
; RV32-NEXT: seqz a1, a1
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: or a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utesth_f16i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: bltu a0, a1, .LBB34_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB34_2: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i64
|
|
%spec.store.select = call i64 @llvm.umin.i64(i64 %conv, i64 4294967295)
|
|
%conv6 = trunc i64 %spec.store.select to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
define i32 @ustest_f16i32_mm(half %x) {
|
|
; RV32-LABEL: ustest_f16i32_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: call __fixsfdi
|
|
; RV32-NEXT: bnez a1, .LBB35_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: li a2, 1
|
|
; RV32-NEXT: j .LBB35_3
|
|
; RV32-NEXT: .LBB35_2:
|
|
; RV32-NEXT: slti a2, a1, 1
|
|
; RV32-NEXT: .LBB35_3: # %entry
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: neg a2, a2
|
|
; RV32-NEXT: or a0, a3, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: srli a1, a1, 31
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i32_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: li a1, -1
|
|
; RV64-NEXT: srli a1, a1, 32
|
|
; RV64-NEXT: blt a0, a1, .LBB35_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB35_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i64
|
|
%spec.store.select = call i64 @llvm.smin.i64(i64 %conv, i64 4294967295)
|
|
%spec.store.select7 = call i64 @llvm.smax.i64(i64 %spec.store.select, i64 0)
|
|
%conv6 = trunc i64 %spec.store.select7 to i32
|
|
ret i32 %conv6
|
|
}
|
|
|
|
; i16 saturate
|
|
|
|
define i16 @stest_f64i16_mm(double %x) {
|
|
; RV32IF-LABEL: stest_f64i16_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfsi
|
|
; RV32IF-NEXT: lui a1, 8
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: blt a0, a1, .LBB36_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB36_2: # %entry
|
|
; RV32IF-NEXT: lui a1, 1048568
|
|
; RV32IF-NEXT: blt a1, a0, .LBB36_4
|
|
; RV32IF-NEXT: # %bb.3: # %entry
|
|
; RV32IF-NEXT: lui a0, 1048568
|
|
; RV32IF-NEXT: .LBB36_4: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i16_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfsi
|
|
; RV64IF-NEXT: lui a1, 8
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: blt a0, a1, .LBB36_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB36_2: # %entry
|
|
; RV64IF-NEXT: lui a1, 1048568
|
|
; RV64IF-NEXT: blt a1, a0, .LBB36_4
|
|
; RV64IF-NEXT: # %bb.3: # %entry
|
|
; RV64IF-NEXT: lui a0, 1048568
|
|
; RV64IF-NEXT: .LBB36_4: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i16_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 8
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: bge a0, a1, .LBB36_3
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: lui a1, 1048568
|
|
; RV32IFD-NEXT: bge a1, a0, .LBB36_4
|
|
; RV32IFD-NEXT: .LBB36_2: # %entry
|
|
; RV32IFD-NEXT: ret
|
|
; RV32IFD-NEXT: .LBB36_3: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: lui a1, 1048568
|
|
; RV32IFD-NEXT: blt a1, a0, .LBB36_2
|
|
; RV32IFD-NEXT: .LBB36_4: # %entry
|
|
; RV32IFD-NEXT: lui a0, 1048568
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i16_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 8
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: bge a0, a1, .LBB36_3
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: lui a1, 1048568
|
|
; RV64IFD-NEXT: bge a1, a0, .LBB36_4
|
|
; RV64IFD-NEXT: .LBB36_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
; RV64IFD-NEXT: .LBB36_3: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: lui a1, 1048568
|
|
; RV64IFD-NEXT: blt a1, a0, .LBB36_2
|
|
; RV64IFD-NEXT: .LBB36_4: # %entry
|
|
; RV64IFD-NEXT: lui a0, 1048568
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 32767)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 -32768)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utest_f64i16_mm(double %x) {
|
|
; RV32IF-LABEL: utest_f64i16_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixunsdfsi
|
|
; RV32IF-NEXT: lui a1, 16
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: bltu a0, a1, .LBB37_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB37_2: # %entry
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: utest_f64i16_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixunsdfsi
|
|
; RV64IF-NEXT: lui a1, 16
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: bltu a0, a1, .LBB37_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB37_2: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i16_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 16
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: bltu a0, a1, .LBB37_2
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: .LBB37_2: # %entry
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: utest_f64i16_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.wu.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 16
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: bltu a0, a1, .LBB37_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB37_2: # %entry
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i32
|
|
%spec.store.select = call i32 @llvm.umin.i32(i32 %conv, i32 65535)
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f64i16_mm(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i16_mm:
|
|
; RV32IF: # %bb.0: # %entry
|
|
; RV32IF-NEXT: addi sp, sp, -16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: call __fixdfsi
|
|
; RV32IF-NEXT: lui a1, 16
|
|
; RV32IF-NEXT: addi a1, a1, -1
|
|
; RV32IF-NEXT: blt a0, a1, .LBB38_2
|
|
; RV32IF-NEXT: # %bb.1: # %entry
|
|
; RV32IF-NEXT: mv a0, a1
|
|
; RV32IF-NEXT: .LBB38_2: # %entry
|
|
; RV32IF-NEXT: sgtz a1, a0
|
|
; RV32IF-NEXT: neg a1, a1
|
|
; RV32IF-NEXT: and a0, a1, a0
|
|
; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: addi sp, sp, 16
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: ustest_f64i16_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfsi
|
|
; RV64IF-NEXT: lui a1, 16
|
|
; RV64IF-NEXT: addi a1, a1, -1
|
|
; RV64IF-NEXT: blt a0, a1, .LBB38_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB38_2: # %entry
|
|
; RV64IF-NEXT: sgtz a1, a0
|
|
; RV64IF-NEXT: neg a1, a1
|
|
; RV64IF-NEXT: and a0, a1, a0
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i16_mm:
|
|
; RV32IFD: # %bb.0: # %entry
|
|
; RV32IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV32IFD-NEXT: lui a1, 16
|
|
; RV32IFD-NEXT: addi a1, a1, -1
|
|
; RV32IFD-NEXT: blt a0, a1, .LBB38_2
|
|
; RV32IFD-NEXT: # %bb.1: # %entry
|
|
; RV32IFD-NEXT: mv a0, a1
|
|
; RV32IFD-NEXT: .LBB38_2: # %entry
|
|
; RV32IFD-NEXT: sgtz a1, a0
|
|
; RV32IFD-NEXT: neg a1, a1
|
|
; RV32IFD-NEXT: and a0, a1, a0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: ustest_f64i16_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.w.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: lui a1, 16
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: blt a0, a1, .LBB38_2
|
|
; RV64IFD-NEXT: # %bb.1: # %entry
|
|
; RV64IFD-NEXT: mv a0, a1
|
|
; RV64IFD-NEXT: .LBB38_2: # %entry
|
|
; RV64IFD-NEXT: sgtz a1, a0
|
|
; RV64IFD-NEXT: neg a1, a1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 65535)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 0)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @stest_f32i16_mm(float %x) {
|
|
; RV32-LABEL: stest_f32i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 8
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bge a0, a1, .LBB39_3
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: bge a1, a0, .LBB39_4
|
|
; RV32-NEXT: .LBB39_2: # %entry
|
|
; RV32-NEXT: ret
|
|
; RV32-NEXT: .LBB39_3: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: blt a1, a0, .LBB39_2
|
|
; RV32-NEXT: .LBB39_4: # %entry
|
|
; RV32-NEXT: lui a0, 1048568
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 8
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bge a0, a1, .LBB39_3
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: bge a1, a0, .LBB39_4
|
|
; RV64-NEXT: .LBB39_2: # %entry
|
|
; RV64-NEXT: ret
|
|
; RV64-NEXT: .LBB39_3: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: blt a1, a0, .LBB39_2
|
|
; RV64-NEXT: .LBB39_4: # %entry
|
|
; RV64-NEXT: lui a0, 1048568
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 32767)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 -32768)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utest_f32i16_mm(float %x) {
|
|
; RV32-LABEL: utest_f32i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bltu a0, a1, .LBB40_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB40_2: # %entry
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bltu a0, a1, .LBB40_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB40_2: # %entry
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i32
|
|
%spec.store.select = call i32 @llvm.umin.i32(i32 %conv, i32 65535)
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f32i16_mm(float %x) {
|
|
; RV32-LABEL: ustest_f32i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB41_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB41_2: # %entry
|
|
; RV32-NEXT: sgtz a1, a0
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB41_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB41_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 65535)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 0)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @stest_f16i16_mm(half %x) {
|
|
; RV32-LABEL: stest_f16i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 8
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB42_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB42_2: # %entry
|
|
; RV32-NEXT: lui a1, 1048568
|
|
; RV32-NEXT: blt a1, a0, .LBB42_4
|
|
; RV32-NEXT: # %bb.3: # %entry
|
|
; RV32-NEXT: lui a0, 1048568
|
|
; RV32-NEXT: .LBB42_4: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 8
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB42_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB42_2: # %entry
|
|
; RV64-NEXT: lui a1, 1048568
|
|
; RV64-NEXT: blt a1, a0, .LBB42_4
|
|
; RV64-NEXT: # %bb.3: # %entry
|
|
; RV64-NEXT: lui a0, 1048568
|
|
; RV64-NEXT: .LBB42_4: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 32767)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 -32768)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @utesth_f16i16_mm(half %x) {
|
|
; RV32-LABEL: utesth_f16i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: bltu a0, a1, .LBB43_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB43_2: # %entry
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utesth_f16i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.lu.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: bltu a0, a1, .LBB43_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB43_2: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i32
|
|
%spec.store.select = call i32 @llvm.umin.i32(i32 %conv, i32 65535)
|
|
%conv6 = trunc i32 %spec.store.select to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
define i16 @ustest_f16i16_mm(half %x) {
|
|
; RV32-LABEL: ustest_f16i16_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: lui a1, 16
|
|
; RV32-NEXT: addi a1, a1, -1
|
|
; RV32-NEXT: blt a0, a1, .LBB44_2
|
|
; RV32-NEXT: # %bb.1: # %entry
|
|
; RV32-NEXT: mv a0, a1
|
|
; RV32-NEXT: .LBB44_2: # %entry
|
|
; RV32-NEXT: sgtz a1, a0
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i16_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: lui a1, 16
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: blt a0, a1, .LBB44_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB44_2: # %entry
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 %conv, i32 65535)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 0)
|
|
%conv6 = trunc i32 %spec.store.select7 to i16
|
|
ret i16 %conv6
|
|
}
|
|
|
|
; i64 saturate
|
|
|
|
define i64 @stest_f64i64_mm(double %x) {
|
|
; RV32IF-LABEL: stest_f64i64_mm:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: addi sp, sp, -80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32IF-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s9, 36(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: .cfi_offset s0, -8
|
|
; RV32IF-NEXT: .cfi_offset s1, -12
|
|
; RV32IF-NEXT: .cfi_offset s2, -16
|
|
; RV32IF-NEXT: .cfi_offset s3, -20
|
|
; RV32IF-NEXT: .cfi_offset s4, -24
|
|
; RV32IF-NEXT: .cfi_offset s5, -28
|
|
; RV32IF-NEXT: .cfi_offset s6, -32
|
|
; RV32IF-NEXT: .cfi_offset s7, -36
|
|
; RV32IF-NEXT: .cfi_offset s8, -40
|
|
; RV32IF-NEXT: .cfi_offset s9, -44
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB45_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li s1, 0
|
|
; RV32IF-NEXT: li s2, 0
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: j .LBB45_8
|
|
; RV32IF-NEXT: .LBB45_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: srai s0, a1, 31
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: ori s2, s0, 1
|
|
; RV32IF-NEXT: addi a5, a3, -1
|
|
; RV32IF-NEXT: addi a3, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a5
|
|
; RV32IF-NEXT: add a0, a1, a4
|
|
; RV32IF-NEXT: seqz a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a5
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a0, a0, a1
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB45_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a1, 1043
|
|
; RV32IF-NEXT: li a4, 1075
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: sub a2, a4, a2
|
|
; RV32IF-NEXT: srai a4, a1, 31
|
|
; RV32IF-NEXT: srl a5, a0, a2
|
|
; RV32IF-NEXT: and s3, a4, a5
|
|
; RV32IF-NEXT: bltz a1, .LBB45_6
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl s4, a0, a1
|
|
; RV32IF-NEXT: j .LBB45_7
|
|
; RV32IF-NEXT: .LBB45_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: sw a3, 16(sp)
|
|
; RV32IF-NEXT: sw a0, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 0(a1)
|
|
; RV32IF-NEXT: lw a2, 4(a1)
|
|
; RV32IF-NEXT: lw a5, 8(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a6, a4, 1
|
|
; RV32IF-NEXT: srli a7, a2, 1
|
|
; RV32IF-NEXT: srli t0, a5, 1
|
|
; RV32IF-NEXT: srl a6, a6, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl a3, t0, a3
|
|
; RV32IF-NEXT: sll a2, a2, a0
|
|
; RV32IF-NEXT: sll a5, a5, a0
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or s3, a2, a6
|
|
; RV32IF-NEXT: or a2, a5, a7
|
|
; RV32IF-NEXT: or a3, a1, a3
|
|
; RV32IF-NEXT: sll s4, a4, a0
|
|
; RV32IF-NEXT: mv a0, s2
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s7, a0, s1
|
|
; RV32IF-NEXT: sltu a0, s7, a0
|
|
; RV32IF-NEXT: add s8, a1, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: add s2, s5, a1
|
|
; RV32IF-NEXT: sltu a0, s2, s5
|
|
; RV32IF-NEXT: add s6, s6, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s4, s6, a0
|
|
; RV32IF-NEXT: sltu s5, s4, s6
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s4, a0, s4
|
|
; RV32IF-NEXT: sltu a2, s4, a0
|
|
; RV32IF-NEXT: add a0, s4, s7
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: add a1, a1, s8
|
|
; RV32IF-NEXT: sltu a2, a0, s4
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: j .LBB45_8
|
|
; RV32IF-NEXT: .LBB45_6:
|
|
; RV32IF-NEXT: srl a1, a3, a2
|
|
; RV32IF-NEXT: not a2, a2
|
|
; RV32IF-NEXT: slli a0, a0, 1
|
|
; RV32IF-NEXT: sll a0, a0, a2
|
|
; RV32IF-NEXT: or s4, a1, a0
|
|
; RV32IF-NEXT: .LBB45_7: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s2
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add s7, a1, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s2, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s2, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s8, s7, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: add s9, a0, s8
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a0, s9, a0
|
|
; RV32IF-NEXT: sltu a2, s9, s5
|
|
; RV32IF-NEXT: sltu a3, s8, s7
|
|
; RV32IF-NEXT: sltu a4, a0, s9
|
|
; RV32IF-NEXT: add a3, s6, a3
|
|
; RV32IF-NEXT: add a2, a3, a2
|
|
; RV32IF-NEXT: add a1, a2, a1
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: .LBB45_8: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: lui a2, 524288
|
|
; RV32IF-NEXT: addi a3, a2, -1
|
|
; RV32IF-NEXT: beq s2, a3, .LBB45_10
|
|
; RV32IF-NEXT: # %bb.9: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: sltu a4, s2, a3
|
|
; RV32IF-NEXT: or a5, a0, a1
|
|
; RV32IF-NEXT: bnez a5, .LBB45_11
|
|
; RV32IF-NEXT: j .LBB45_12
|
|
; RV32IF-NEXT: .LBB45_10:
|
|
; RV32IF-NEXT: sltiu a4, s1, -1
|
|
; RV32IF-NEXT: or a5, a0, a1
|
|
; RV32IF-NEXT: beqz a5, .LBB45_12
|
|
; RV32IF-NEXT: .LBB45_11: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a4, a1, 31
|
|
; RV32IF-NEXT: .LBB45_12: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: neg a5, a4
|
|
; RV32IF-NEXT: addi a6, a4, -1
|
|
; RV32IF-NEXT: bnez a4, .LBB45_14
|
|
; RV32IF-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: mv s2, a3
|
|
; RV32IF-NEXT: .LBB45_14: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: or a3, a6, s1
|
|
; RV32IF-NEXT: and a1, a5, a1
|
|
; RV32IF-NEXT: and a4, a5, a0
|
|
; RV32IF-NEXT: beq s2, a2, .LBB45_16
|
|
; RV32IF-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: sltu a0, a2, s2
|
|
; RV32IF-NEXT: j .LBB45_17
|
|
; RV32IF-NEXT: .LBB45_16:
|
|
; RV32IF-NEXT: snez a0, a3
|
|
; RV32IF-NEXT: .LBB45_17: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: and a4, a4, a1
|
|
; RV32IF-NEXT: li a2, -1
|
|
; RV32IF-NEXT: beq a4, a2, .LBB45_19
|
|
; RV32IF-NEXT: # %bb.18: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a1, a1, 31
|
|
; RV32IF-NEXT: xori a0, a1, 1
|
|
; RV32IF-NEXT: .LBB45_19: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: bnez a0, .LBB45_21
|
|
; RV32IF-NEXT: # %bb.20: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: lui s2, 524288
|
|
; RV32IF-NEXT: .LBB45_21: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: neg a0, a0
|
|
; RV32IF-NEXT: and a0, a0, a3
|
|
; RV32IF-NEXT: mv a1, s2
|
|
; RV32IF-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s9, 36(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: .cfi_restore s0
|
|
; RV32IF-NEXT: .cfi_restore s1
|
|
; RV32IF-NEXT: .cfi_restore s2
|
|
; RV32IF-NEXT: .cfi_restore s3
|
|
; RV32IF-NEXT: .cfi_restore s4
|
|
; RV32IF-NEXT: .cfi_restore s5
|
|
; RV32IF-NEXT: .cfi_restore s6
|
|
; RV32IF-NEXT: .cfi_restore s7
|
|
; RV32IF-NEXT: .cfi_restore s8
|
|
; RV32IF-NEXT: .cfi_restore s9
|
|
; RV32IF-NEXT: addi sp, sp, 80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64IF-LABEL: stest_f64i64_mm:
|
|
; RV64IF: # %bb.0: # %entry
|
|
; RV64IF-NEXT: addi sp, sp, -16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64IF-NEXT: .cfi_offset ra, -8
|
|
; RV64IF-NEXT: call __fixdfti
|
|
; RV64IF-NEXT: li a2, -1
|
|
; RV64IF-NEXT: srli a3, a2, 1
|
|
; RV64IF-NEXT: beqz a1, .LBB45_2
|
|
; RV64IF-NEXT: # %bb.1: # %entry
|
|
; RV64IF-NEXT: srli a4, a1, 63
|
|
; RV64IF-NEXT: j .LBB45_3
|
|
; RV64IF-NEXT: .LBB45_2:
|
|
; RV64IF-NEXT: sltu a4, a0, a3
|
|
; RV64IF-NEXT: .LBB45_3: # %entry
|
|
; RV64IF-NEXT: neg a5, a4
|
|
; RV64IF-NEXT: and a5, a5, a1
|
|
; RV64IF-NEXT: bnez a4, .LBB45_5
|
|
; RV64IF-NEXT: # %bb.4: # %entry
|
|
; RV64IF-NEXT: mv a0, a3
|
|
; RV64IF-NEXT: .LBB45_5: # %entry
|
|
; RV64IF-NEXT: slli a1, a2, 63
|
|
; RV64IF-NEXT: beq a5, a2, .LBB45_7
|
|
; RV64IF-NEXT: # %bb.6: # %entry
|
|
; RV64IF-NEXT: srli a5, a5, 63
|
|
; RV64IF-NEXT: xori a2, a5, 1
|
|
; RV64IF-NEXT: beqz a2, .LBB45_8
|
|
; RV64IF-NEXT: j .LBB45_9
|
|
; RV64IF-NEXT: .LBB45_7:
|
|
; RV64IF-NEXT: sltu a2, a1, a0
|
|
; RV64IF-NEXT: bnez a2, .LBB45_9
|
|
; RV64IF-NEXT: .LBB45_8: # %entry
|
|
; RV64IF-NEXT: mv a0, a1
|
|
; RV64IF-NEXT: .LBB45_9: # %entry
|
|
; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64IF-NEXT: .cfi_restore ra
|
|
; RV64IF-NEXT: addi sp, sp, 16
|
|
; RV64IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64IF-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: stest_f64i64_mm:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 96
|
|
; RV32IFD-NEXT: sw ra, 92(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s0, 88(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s1, 84(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s2, 80(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s3, 76(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s4, 72(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s5, 68(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s6, 64(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s7, 60(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s8, 56(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s9, 52(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: .cfi_offset ra, -4
|
|
; RV32IFD-NEXT: .cfi_offset s0, -8
|
|
; RV32IFD-NEXT: .cfi_offset s1, -12
|
|
; RV32IFD-NEXT: .cfi_offset s2, -16
|
|
; RV32IFD-NEXT: .cfi_offset s3, -20
|
|
; RV32IFD-NEXT: .cfi_offset s4, -24
|
|
; RV32IFD-NEXT: .cfi_offset s5, -28
|
|
; RV32IFD-NEXT: .cfi_offset s6, -32
|
|
; RV32IFD-NEXT: .cfi_offset s7, -36
|
|
; RV32IFD-NEXT: .cfi_offset s8, -40
|
|
; RV32IFD-NEXT: .cfi_offset s9, -44
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB45_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li s1, 0
|
|
; RV32IFD-NEXT: li s2, 0
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: j .LBB45_8
|
|
; RV32IFD-NEXT: .LBB45_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: srai s0, a1, 31
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: ori s2, s0, 1
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB45_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 1043
|
|
; RV32IFD-NEXT: li a4, 1075
|
|
; RV32IFD-NEXT: sub a3, a3, a0
|
|
; RV32IFD-NEXT: sub a0, a4, a0
|
|
; RV32IFD-NEXT: srai a4, a3, 31
|
|
; RV32IFD-NEXT: srl a5, a1, a0
|
|
; RV32IFD-NEXT: and s3, a4, a5
|
|
; RV32IFD-NEXT: bltz a3, .LBB45_6
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl s4, a1, a3
|
|
; RV32IFD-NEXT: j .LBB45_7
|
|
; RV32IFD-NEXT: .LBB45_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 0(a1)
|
|
; RV32IFD-NEXT: lw a2, 4(a1)
|
|
; RV32IFD-NEXT: lw a5, 8(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a6, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a2, 1
|
|
; RV32IFD-NEXT: srli t0, a5, 1
|
|
; RV32IFD-NEXT: srl a6, a6, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl a3, t0, a3
|
|
; RV32IFD-NEXT: sll a2, a2, a0
|
|
; RV32IFD-NEXT: sll a5, a5, a0
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or s3, a2, a6
|
|
; RV32IFD-NEXT: or a2, a5, a7
|
|
; RV32IFD-NEXT: or a3, a1, a3
|
|
; RV32IFD-NEXT: sll s4, a4, a0
|
|
; RV32IFD-NEXT: mv a0, s2
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s7, a0, s1
|
|
; RV32IFD-NEXT: sltu a0, s7, a0
|
|
; RV32IFD-NEXT: add s8, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: add s2, s5, a1
|
|
; RV32IFD-NEXT: sltu a0, s2, s5
|
|
; RV32IFD-NEXT: add s6, s6, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s4, s6, a0
|
|
; RV32IFD-NEXT: sltu s5, s4, s6
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s4, a0, s4
|
|
; RV32IFD-NEXT: sltu a2, s4, a0
|
|
; RV32IFD-NEXT: add a0, s4, s7
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: add a1, a1, s8
|
|
; RV32IFD-NEXT: sltu a2, a0, s4
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: j .LBB45_8
|
|
; RV32IFD-NEXT: .LBB45_6:
|
|
; RV32IFD-NEXT: srl a2, a2, a0
|
|
; RV32IFD-NEXT: not a0, a0
|
|
; RV32IFD-NEXT: slli a1, a1, 1
|
|
; RV32IFD-NEXT: sll a0, a1, a0
|
|
; RV32IFD-NEXT: or s4, a2, a0
|
|
; RV32IFD-NEXT: .LBB45_7: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s2
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add s7, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s2, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s2, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s8, s7, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: add s9, a0, s8
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a0, s9, a0
|
|
; RV32IFD-NEXT: sltu a2, s9, s5
|
|
; RV32IFD-NEXT: sltu a3, s8, s7
|
|
; RV32IFD-NEXT: sltu a4, a0, s9
|
|
; RV32IFD-NEXT: add a3, s6, a3
|
|
; RV32IFD-NEXT: add a2, a3, a2
|
|
; RV32IFD-NEXT: add a1, a2, a1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: .LBB45_8: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: lui a2, 524288
|
|
; RV32IFD-NEXT: addi a3, a2, -1
|
|
; RV32IFD-NEXT: beq s2, a3, .LBB45_10
|
|
; RV32IFD-NEXT: # %bb.9: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: sltu a4, s2, a3
|
|
; RV32IFD-NEXT: or a5, a0, a1
|
|
; RV32IFD-NEXT: bnez a5, .LBB45_11
|
|
; RV32IFD-NEXT: j .LBB45_12
|
|
; RV32IFD-NEXT: .LBB45_10:
|
|
; RV32IFD-NEXT: sltiu a4, s1, -1
|
|
; RV32IFD-NEXT: or a5, a0, a1
|
|
; RV32IFD-NEXT: beqz a5, .LBB45_12
|
|
; RV32IFD-NEXT: .LBB45_11: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a4, a1, 31
|
|
; RV32IFD-NEXT: .LBB45_12: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: neg a5, a4
|
|
; RV32IFD-NEXT: addi a6, a4, -1
|
|
; RV32IFD-NEXT: bnez a4, .LBB45_14
|
|
; RV32IFD-NEXT: # %bb.13: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: mv s2, a3
|
|
; RV32IFD-NEXT: .LBB45_14: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: or a3, a6, s1
|
|
; RV32IFD-NEXT: and a1, a5, a1
|
|
; RV32IFD-NEXT: and a4, a5, a0
|
|
; RV32IFD-NEXT: beq s2, a2, .LBB45_16
|
|
; RV32IFD-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: sltu a0, a2, s2
|
|
; RV32IFD-NEXT: j .LBB45_17
|
|
; RV32IFD-NEXT: .LBB45_16:
|
|
; RV32IFD-NEXT: snez a0, a3
|
|
; RV32IFD-NEXT: .LBB45_17: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: and a4, a4, a1
|
|
; RV32IFD-NEXT: li a2, -1
|
|
; RV32IFD-NEXT: beq a4, a2, .LBB45_19
|
|
; RV32IFD-NEXT: # %bb.18: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a1, a1, 31
|
|
; RV32IFD-NEXT: xori a0, a1, 1
|
|
; RV32IFD-NEXT: .LBB45_19: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: bnez a0, .LBB45_21
|
|
; RV32IFD-NEXT: # %bb.20: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: lui s2, 524288
|
|
; RV32IFD-NEXT: .LBB45_21: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: neg a0, a0
|
|
; RV32IFD-NEXT: and a0, a0, a3
|
|
; RV32IFD-NEXT: mv a1, s2
|
|
; RV32IFD-NEXT: lw ra, 92(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s0, 88(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s1, 84(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s2, 80(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s3, 76(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s4, 72(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s5, 68(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s6, 64(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s7, 60(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s8, 56(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s9, 52(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: .cfi_restore ra
|
|
; RV32IFD-NEXT: .cfi_restore s0
|
|
; RV32IFD-NEXT: .cfi_restore s1
|
|
; RV32IFD-NEXT: .cfi_restore s2
|
|
; RV32IFD-NEXT: .cfi_restore s3
|
|
; RV32IFD-NEXT: .cfi_restore s4
|
|
; RV32IFD-NEXT: .cfi_restore s5
|
|
; RV32IFD-NEXT: .cfi_restore s6
|
|
; RV32IFD-NEXT: .cfi_restore s7
|
|
; RV32IFD-NEXT: .cfi_restore s8
|
|
; RV32IFD-NEXT: .cfi_restore s9
|
|
; RV32IFD-NEXT: addi sp, sp, 96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
;
|
|
; RV64IFD-LABEL: stest_f64i64_mm:
|
|
; RV64IFD: # %bb.0: # %entry
|
|
; RV64IFD-NEXT: fcvt.l.d a0, fa0, rtz
|
|
; RV64IFD-NEXT: feq.d a1, fa0, fa0
|
|
; RV64IFD-NEXT: seqz a1, a1
|
|
; RV64IFD-NEXT: addi a1, a1, -1
|
|
; RV64IFD-NEXT: and a0, a1, a0
|
|
; RV64IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 9223372036854775807)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 -9223372036854775808)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utest_f64i64_mm(double %x) {
|
|
; RV32IF-LABEL: utest_f64i64_mm:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB46_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: li a2, 0
|
|
; RV32IF-NEXT: j .LBB46_8
|
|
; RV32IF-NEXT: .LBB46_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: addi a3, a3, -1
|
|
; RV32IF-NEXT: addi a0, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: seqz a4, a0
|
|
; RV32IF-NEXT: add a3, a4, a3
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a1, a1, a3
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB46_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a5, 1075
|
|
; RV32IF-NEXT: li a4, 1043
|
|
; RV32IF-NEXT: sub a4, a4, a2
|
|
; RV32IF-NEXT: sub a5, a5, a2
|
|
; RV32IF-NEXT: bltz a4, .LBB46_6
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl a0, a1, a4
|
|
; RV32IF-NEXT: j .LBB46_7
|
|
; RV32IF-NEXT: .LBB46_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: addi sp, sp, -32
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 32
|
|
; RV32IF-NEXT: sw a0, 16(sp)
|
|
; RV32IF-NEXT: sw a1, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 8(a1)
|
|
; RV32IF-NEXT: lw a5, 4(a1)
|
|
; RV32IF-NEXT: lw a6, 0(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a2, a4, 1
|
|
; RV32IF-NEXT: srli a7, a5, 1
|
|
; RV32IF-NEXT: srli t0, a6, 1
|
|
; RV32IF-NEXT: srl a2, a2, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl t0, t0, a3
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or a2, a1, a2
|
|
; RV32IF-NEXT: sll a1, a4, a0
|
|
; RV32IF-NEXT: sll a4, a5, a0
|
|
; RV32IF-NEXT: or a3, a1, a7
|
|
; RV32IF-NEXT: or a1, a4, t0
|
|
; RV32IF-NEXT: sll a0, a6, a0
|
|
; RV32IF-NEXT: addi sp, sp, 32
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: j .LBB46_8
|
|
; RV32IF-NEXT: .LBB46_6:
|
|
; RV32IF-NEXT: srl a0, a0, a5
|
|
; RV32IF-NEXT: not a2, a5
|
|
; RV32IF-NEXT: slli a3, a1, 1
|
|
; RV32IF-NEXT: sll a2, a3, a2
|
|
; RV32IF-NEXT: or a0, a0, a2
|
|
; RV32IF-NEXT: .LBB46_7: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: li a2, 0
|
|
; RV32IF-NEXT: srl a1, a1, a5
|
|
; RV32IF-NEXT: srai a4, a4, 31
|
|
; RV32IF-NEXT: and a1, a4, a1
|
|
; RV32IF-NEXT: .LBB46_8: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: or a4, a2, a3
|
|
; RV32IF-NEXT: xori a3, a3, 1
|
|
; RV32IF-NEXT: seqz a4, a4
|
|
; RV32IF-NEXT: or a2, a3, a2
|
|
; RV32IF-NEXT: seqz a2, a2
|
|
; RV32IF-NEXT: addi a2, a2, -1
|
|
; RV32IF-NEXT: and a2, a2, a4
|
|
; RV32IF-NEXT: neg a2, a2
|
|
; RV32IF-NEXT: and a0, a2, a0
|
|
; RV32IF-NEXT: and a1, a2, a1
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f64i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixunsdfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: utest_f64i64_mm:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -48
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 48
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB46_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: li a2, 0
|
|
; RV32IFD-NEXT: j .LBB46_8
|
|
; RV32IFD-NEXT: .LBB46_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB46_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a5, 1075
|
|
; RV32IFD-NEXT: li a4, 1043
|
|
; RV32IFD-NEXT: sub a4, a4, a0
|
|
; RV32IFD-NEXT: sub a5, a5, a0
|
|
; RV32IFD-NEXT: bltz a4, .LBB46_6
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl a0, a1, a4
|
|
; RV32IFD-NEXT: j .LBB46_7
|
|
; RV32IFD-NEXT: .LBB46_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 8(a1)
|
|
; RV32IFD-NEXT: lw a5, 4(a1)
|
|
; RV32IFD-NEXT: lw a6, 0(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a2, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a5, 1
|
|
; RV32IFD-NEXT: srli t0, a6, 1
|
|
; RV32IFD-NEXT: srl a2, a2, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl t0, t0, a3
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or a2, a1, a2
|
|
; RV32IFD-NEXT: sll a1, a4, a0
|
|
; RV32IFD-NEXT: sll a4, a5, a0
|
|
; RV32IFD-NEXT: or a3, a1, a7
|
|
; RV32IFD-NEXT: or a1, a4, t0
|
|
; RV32IFD-NEXT: sll a0, a6, a0
|
|
; RV32IFD-NEXT: j .LBB46_8
|
|
; RV32IFD-NEXT: .LBB46_6:
|
|
; RV32IFD-NEXT: srl a0, a2, a5
|
|
; RV32IFD-NEXT: not a2, a5
|
|
; RV32IFD-NEXT: slli a3, a1, 1
|
|
; RV32IFD-NEXT: sll a2, a3, a2
|
|
; RV32IFD-NEXT: or a0, a0, a2
|
|
; RV32IFD-NEXT: .LBB46_7: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: li a2, 0
|
|
; RV32IFD-NEXT: srl a1, a1, a5
|
|
; RV32IFD-NEXT: srai a4, a4, 31
|
|
; RV32IFD-NEXT: and a1, a4, a1
|
|
; RV32IFD-NEXT: .LBB46_8: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: or a4, a2, a3
|
|
; RV32IFD-NEXT: xori a3, a3, 1
|
|
; RV32IFD-NEXT: seqz a4, a4
|
|
; RV32IFD-NEXT: or a2, a3, a2
|
|
; RV32IFD-NEXT: seqz a2, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: and a2, a2, a4
|
|
; RV32IFD-NEXT: neg a2, a2
|
|
; RV32IFD-NEXT: and a0, a2, a0
|
|
; RV32IFD-NEXT: and a1, a2, a1
|
|
; RV32IFD-NEXT: addi sp, sp, 48
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptoui double %x to i128
|
|
%spec.store.select = call i128 @llvm.umin.i128(i128 %conv, i128 18446744073709551616)
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f64i64_mm(double %x) {
|
|
; RV32IF-LABEL: ustest_f64i64_mm:
|
|
; RV32IF: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IF-NEXT: addi sp, sp, -80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32IF-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: sw s9, 36(sp) # 4-byte Folded Spill
|
|
; RV32IF-NEXT: .cfi_offset ra, -4
|
|
; RV32IF-NEXT: .cfi_offset s0, -8
|
|
; RV32IF-NEXT: .cfi_offset s1, -12
|
|
; RV32IF-NEXT: .cfi_offset s2, -16
|
|
; RV32IF-NEXT: .cfi_offset s3, -20
|
|
; RV32IF-NEXT: .cfi_offset s4, -24
|
|
; RV32IF-NEXT: .cfi_offset s5, -28
|
|
; RV32IF-NEXT: .cfi_offset s6, -32
|
|
; RV32IF-NEXT: .cfi_offset s7, -36
|
|
; RV32IF-NEXT: .cfi_offset s8, -40
|
|
; RV32IF-NEXT: .cfi_offset s9, -44
|
|
; RV32IF-NEXT: slli a2, a1, 1
|
|
; RV32IF-NEXT: srli a2, a2, 21
|
|
; RV32IF-NEXT: li a3, 1023
|
|
; RV32IF-NEXT: bgeu a2, a3, .LBB47_2
|
|
; RV32IF-NEXT: # %bb.1:
|
|
; RV32IF-NEXT: li s2, 0
|
|
; RV32IF-NEXT: li s6, 0
|
|
; RV32IF-NEXT: li a0, 0
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: j .LBB47_6
|
|
; RV32IF-NEXT: .LBB47_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IF-NEXT: srai s0, a1, 31
|
|
; RV32IF-NEXT: lui a3, 256
|
|
; RV32IF-NEXT: snez a4, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: ori s1, s0, 1
|
|
; RV32IF-NEXT: addi a5, a3, -1
|
|
; RV32IF-NEXT: addi a3, a0, 1
|
|
; RV32IF-NEXT: and a1, a1, a5
|
|
; RV32IF-NEXT: add a0, a1, a4
|
|
; RV32IF-NEXT: seqz a1, a3
|
|
; RV32IF-NEXT: add a1, a1, a5
|
|
; RV32IF-NEXT: li a4, 1074
|
|
; RV32IF-NEXT: add a0, a0, a1
|
|
; RV32IF-NEXT: bltu a4, a2, .LBB47_5
|
|
; RV32IF-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: li a1, 1043
|
|
; RV32IF-NEXT: li a4, 1075
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: sub a2, a4, a2
|
|
; RV32IF-NEXT: srai a4, a1, 31
|
|
; RV32IF-NEXT: srl a5, a0, a2
|
|
; RV32IF-NEXT: and s3, a4, a5
|
|
; RV32IF-NEXT: bltz a1, .LBB47_7
|
|
; RV32IF-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: srl s4, a0, a1
|
|
; RV32IF-NEXT: j .LBB47_8
|
|
; RV32IF-NEXT: .LBB47_5: # %fp-to-i-if-exp.large
|
|
; RV32IF-NEXT: sw a3, 16(sp)
|
|
; RV32IF-NEXT: sw a0, 20(sp)
|
|
; RV32IF-NEXT: sw zero, 24(sp)
|
|
; RV32IF-NEXT: sw zero, 28(sp)
|
|
; RV32IF-NEXT: sw zero, 0(sp)
|
|
; RV32IF-NEXT: sw zero, 4(sp)
|
|
; RV32IF-NEXT: sw zero, 8(sp)
|
|
; RV32IF-NEXT: sw zero, 12(sp)
|
|
; RV32IF-NEXT: addi a0, a2, -1075
|
|
; RV32IF-NEXT: addi a1, sp, 16
|
|
; RV32IF-NEXT: srli a2, a0, 3
|
|
; RV32IF-NEXT: andi a3, a0, 31
|
|
; RV32IF-NEXT: andi a2, a2, 12
|
|
; RV32IF-NEXT: xori a3, a3, 31
|
|
; RV32IF-NEXT: sub a1, a1, a2
|
|
; RV32IF-NEXT: lw a4, 0(a1)
|
|
; RV32IF-NEXT: lw a2, 4(a1)
|
|
; RV32IF-NEXT: lw a5, 8(a1)
|
|
; RV32IF-NEXT: lw a1, 12(a1)
|
|
; RV32IF-NEXT: srli a6, a4, 1
|
|
; RV32IF-NEXT: srli a7, a2, 1
|
|
; RV32IF-NEXT: srli t0, a5, 1
|
|
; RV32IF-NEXT: srl a6, a6, a3
|
|
; RV32IF-NEXT: srl a7, a7, a3
|
|
; RV32IF-NEXT: srl a3, t0, a3
|
|
; RV32IF-NEXT: sll a2, a2, a0
|
|
; RV32IF-NEXT: sll a5, a5, a0
|
|
; RV32IF-NEXT: sll a1, a1, a0
|
|
; RV32IF-NEXT: or s3, a2, a6
|
|
; RV32IF-NEXT: or a2, a5, a7
|
|
; RV32IF-NEXT: or a3, a1, a3
|
|
; RV32IF-NEXT: sll s4, a4, a0
|
|
; RV32IF-NEXT: mv a0, s1
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s2, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s5
|
|
; RV32IF-NEXT: add s7, a0, s2
|
|
; RV32IF-NEXT: sltu a0, s7, a0
|
|
; RV32IF-NEXT: add s8, a1, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s1
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s5, a0
|
|
; RV32IF-NEXT: mv s6, a1
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s1
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s2, a0
|
|
; RV32IF-NEXT: add s1, s5, a1
|
|
; RV32IF-NEXT: sltu a0, s1, s5
|
|
; RV32IF-NEXT: add s5, s6, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s6, a0, s1
|
|
; RV32IF-NEXT: sltu a0, s6, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s1, s5, a0
|
|
; RV32IF-NEXT: sltu s4, s1, s5
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a1, a1, s4
|
|
; RV32IF-NEXT: add s1, a0, s1
|
|
; RV32IF-NEXT: sltu a2, s1, a0
|
|
; RV32IF-NEXT: add a0, s1, s7
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: add a1, a1, s8
|
|
; RV32IF-NEXT: sltu a2, a0, s1
|
|
; RV32IF-NEXT: add a1, a1, a2
|
|
; RV32IF-NEXT: bnez a1, .LBB47_9
|
|
; RV32IF-NEXT: .LBB47_6:
|
|
; RV32IF-NEXT: seqz a2, a0
|
|
; RV32IF-NEXT: j .LBB47_10
|
|
; RV32IF-NEXT: .LBB47_7:
|
|
; RV32IF-NEXT: srl a1, a3, a2
|
|
; RV32IF-NEXT: not a2, a2
|
|
; RV32IF-NEXT: slli a0, a0, 1
|
|
; RV32IF-NEXT: sll a0, a0, a2
|
|
; RV32IF-NEXT: or s4, a1, a0
|
|
; RV32IF-NEXT: .LBB47_8: # %fp-to-i-if-exp.small
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s1
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s2, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s1
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s5, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s5, a0
|
|
; RV32IF-NEXT: add s7, a1, a0
|
|
; RV32IF-NEXT: mv a0, s4
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add s6, a0, s5
|
|
; RV32IF-NEXT: sltu a0, s6, a0
|
|
; RV32IF-NEXT: add a0, a1, a0
|
|
; RV32IF-NEXT: add s8, s7, a0
|
|
; RV32IF-NEXT: mv a0, s3
|
|
; RV32IF-NEXT: li a1, 0
|
|
; RV32IF-NEXT: mv a2, s0
|
|
; RV32IF-NEXT: li a3, 0
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: mv s1, a0
|
|
; RV32IF-NEXT: mv s5, a1
|
|
; RV32IF-NEXT: add s9, a0, s8
|
|
; RV32IF-NEXT: mv a0, s0
|
|
; RV32IF-NEXT: mv a1, s0
|
|
; RV32IF-NEXT: mv a2, s4
|
|
; RV32IF-NEXT: mv a3, s3
|
|
; RV32IF-NEXT: call __muldi3
|
|
; RV32IF-NEXT: add a0, s9, a0
|
|
; RV32IF-NEXT: sltu a2, s9, s1
|
|
; RV32IF-NEXT: sltu a3, s8, s7
|
|
; RV32IF-NEXT: sltu a4, a0, s9
|
|
; RV32IF-NEXT: add a3, s5, a3
|
|
; RV32IF-NEXT: add a2, a3, a2
|
|
; RV32IF-NEXT: add a1, a2, a1
|
|
; RV32IF-NEXT: add a1, a1, a4
|
|
; RV32IF-NEXT: beqz a1, .LBB47_6
|
|
; RV32IF-NEXT: .LBB47_9: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: srli a2, a1, 31
|
|
; RV32IF-NEXT: .LBB47_10: # %fp-to-i-cleanup
|
|
; RV32IF-NEXT: xori a0, a0, 1
|
|
; RV32IF-NEXT: or a0, a0, a1
|
|
; RV32IF-NEXT: seqz a0, a0
|
|
; RV32IF-NEXT: addi a0, a0, -1
|
|
; RV32IF-NEXT: and a0, a0, a2
|
|
; RV32IF-NEXT: neg a0, a0
|
|
; RV32IF-NEXT: and a2, a0, s6
|
|
; RV32IF-NEXT: and a3, a0, s2
|
|
; RV32IF-NEXT: and a0, a0, a1
|
|
; RV32IF-NEXT: srli a0, a0, 31
|
|
; RV32IF-NEXT: addi a1, a0, -1
|
|
; RV32IF-NEXT: and a0, a1, a3
|
|
; RV32IF-NEXT: and a1, a1, a2
|
|
; RV32IF-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: lw s9, 36(sp) # 4-byte Folded Reload
|
|
; RV32IF-NEXT: .cfi_restore ra
|
|
; RV32IF-NEXT: .cfi_restore s0
|
|
; RV32IF-NEXT: .cfi_restore s1
|
|
; RV32IF-NEXT: .cfi_restore s2
|
|
; RV32IF-NEXT: .cfi_restore s3
|
|
; RV32IF-NEXT: .cfi_restore s4
|
|
; RV32IF-NEXT: .cfi_restore s5
|
|
; RV32IF-NEXT: .cfi_restore s6
|
|
; RV32IF-NEXT: .cfi_restore s7
|
|
; RV32IF-NEXT: .cfi_restore s8
|
|
; RV32IF-NEXT: .cfi_restore s9
|
|
; RV32IF-NEXT: addi sp, sp, 80
|
|
; RV32IF-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IF-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f64i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixdfti
|
|
; RV64-NEXT: mv a2, a1
|
|
; RV64-NEXT: blez a1, .LBB47_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a2, 1
|
|
; RV64-NEXT: .LBB47_2: # %entry
|
|
; RV64-NEXT: slti a1, a1, 1
|
|
; RV64-NEXT: srli a2, a2, 63
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: addi a2, a2, -1
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
;
|
|
; RV32IFD-LABEL: ustest_f64i64_mm:
|
|
; RV32IFD: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32IFD-NEXT: addi sp, sp, -96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 96
|
|
; RV32IFD-NEXT: sw ra, 92(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s0, 88(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s1, 84(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s2, 80(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s3, 76(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s4, 72(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s5, 68(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s6, 64(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s7, 60(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s8, 56(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: sw s9, 52(sp) # 4-byte Folded Spill
|
|
; RV32IFD-NEXT: .cfi_offset ra, -4
|
|
; RV32IFD-NEXT: .cfi_offset s0, -8
|
|
; RV32IFD-NEXT: .cfi_offset s1, -12
|
|
; RV32IFD-NEXT: .cfi_offset s2, -16
|
|
; RV32IFD-NEXT: .cfi_offset s3, -20
|
|
; RV32IFD-NEXT: .cfi_offset s4, -24
|
|
; RV32IFD-NEXT: .cfi_offset s5, -28
|
|
; RV32IFD-NEXT: .cfi_offset s6, -32
|
|
; RV32IFD-NEXT: .cfi_offset s7, -36
|
|
; RV32IFD-NEXT: .cfi_offset s8, -40
|
|
; RV32IFD-NEXT: .cfi_offset s9, -44
|
|
; RV32IFD-NEXT: fsd fa0, 8(sp)
|
|
; RV32IFD-NEXT: lw a1, 12(sp)
|
|
; RV32IFD-NEXT: slli a0, a1, 1
|
|
; RV32IFD-NEXT: srli a0, a0, 21
|
|
; RV32IFD-NEXT: li a2, 1023
|
|
; RV32IFD-NEXT: bgeu a0, a2, .LBB47_2
|
|
; RV32IFD-NEXT: # %bb.1:
|
|
; RV32IFD-NEXT: li s2, 0
|
|
; RV32IFD-NEXT: li s6, 0
|
|
; RV32IFD-NEXT: li a0, 0
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: j .LBB47_6
|
|
; RV32IFD-NEXT: .LBB47_2: # %fp-to-i-if-check.exp.size
|
|
; RV32IFD-NEXT: lw a2, 8(sp)
|
|
; RV32IFD-NEXT: srai s0, a1, 31
|
|
; RV32IFD-NEXT: lui a3, 256
|
|
; RV32IFD-NEXT: ori s1, s0, 1
|
|
; RV32IFD-NEXT: addi a3, a3, -1
|
|
; RV32IFD-NEXT: and a1, a1, a3
|
|
; RV32IFD-NEXT: snez a4, a2
|
|
; RV32IFD-NEXT: addi a2, a2, -1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: addi a2, a2, 1
|
|
; RV32IFD-NEXT: seqz a4, a2
|
|
; RV32IFD-NEXT: add a1, a1, a3
|
|
; RV32IFD-NEXT: li a3, 1074
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: bltu a3, a0, .LBB47_5
|
|
; RV32IFD-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: li a3, 1043
|
|
; RV32IFD-NEXT: li a4, 1075
|
|
; RV32IFD-NEXT: sub a3, a3, a0
|
|
; RV32IFD-NEXT: sub a0, a4, a0
|
|
; RV32IFD-NEXT: srai a4, a3, 31
|
|
; RV32IFD-NEXT: srl a5, a1, a0
|
|
; RV32IFD-NEXT: and s3, a4, a5
|
|
; RV32IFD-NEXT: bltz a3, .LBB47_7
|
|
; RV32IFD-NEXT: # %bb.4: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: srl s4, a1, a3
|
|
; RV32IFD-NEXT: j .LBB47_8
|
|
; RV32IFD-NEXT: .LBB47_5: # %fp-to-i-if-exp.large
|
|
; RV32IFD-NEXT: sw a2, 32(sp)
|
|
; RV32IFD-NEXT: sw a1, 36(sp)
|
|
; RV32IFD-NEXT: sw zero, 40(sp)
|
|
; RV32IFD-NEXT: sw zero, 44(sp)
|
|
; RV32IFD-NEXT: sw zero, 16(sp)
|
|
; RV32IFD-NEXT: sw zero, 20(sp)
|
|
; RV32IFD-NEXT: sw zero, 24(sp)
|
|
; RV32IFD-NEXT: sw zero, 28(sp)
|
|
; RV32IFD-NEXT: addi a0, a0, -1075
|
|
; RV32IFD-NEXT: addi a1, sp, 32
|
|
; RV32IFD-NEXT: srli a2, a0, 3
|
|
; RV32IFD-NEXT: andi a3, a0, 31
|
|
; RV32IFD-NEXT: andi a2, a2, 12
|
|
; RV32IFD-NEXT: xori a3, a3, 31
|
|
; RV32IFD-NEXT: sub a1, a1, a2
|
|
; RV32IFD-NEXT: lw a4, 0(a1)
|
|
; RV32IFD-NEXT: lw a2, 4(a1)
|
|
; RV32IFD-NEXT: lw a5, 8(a1)
|
|
; RV32IFD-NEXT: lw a1, 12(a1)
|
|
; RV32IFD-NEXT: srli a6, a4, 1
|
|
; RV32IFD-NEXT: srli a7, a2, 1
|
|
; RV32IFD-NEXT: srli t0, a5, 1
|
|
; RV32IFD-NEXT: srl a6, a6, a3
|
|
; RV32IFD-NEXT: srl a7, a7, a3
|
|
; RV32IFD-NEXT: srl a3, t0, a3
|
|
; RV32IFD-NEXT: sll a2, a2, a0
|
|
; RV32IFD-NEXT: sll a5, a5, a0
|
|
; RV32IFD-NEXT: sll a1, a1, a0
|
|
; RV32IFD-NEXT: or s3, a2, a6
|
|
; RV32IFD-NEXT: or a2, a5, a7
|
|
; RV32IFD-NEXT: or a3, a1, a3
|
|
; RV32IFD-NEXT: sll s4, a4, a0
|
|
; RV32IFD-NEXT: mv a0, s1
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s2, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s5
|
|
; RV32IFD-NEXT: add s7, a0, s2
|
|
; RV32IFD-NEXT: sltu a0, s7, a0
|
|
; RV32IFD-NEXT: add s8, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s1
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s5, a0
|
|
; RV32IFD-NEXT: mv s6, a1
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s1
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s2, a0
|
|
; RV32IFD-NEXT: add s1, s5, a1
|
|
; RV32IFD-NEXT: sltu a0, s1, s5
|
|
; RV32IFD-NEXT: add s5, s6, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s6, a0, s1
|
|
; RV32IFD-NEXT: sltu a0, s6, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s1, s5, a0
|
|
; RV32IFD-NEXT: sltu s4, s1, s5
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a1, a1, s4
|
|
; RV32IFD-NEXT: add s1, a0, s1
|
|
; RV32IFD-NEXT: sltu a2, s1, a0
|
|
; RV32IFD-NEXT: add a0, s1, s7
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: add a1, a1, s8
|
|
; RV32IFD-NEXT: sltu a2, a0, s1
|
|
; RV32IFD-NEXT: add a1, a1, a2
|
|
; RV32IFD-NEXT: bnez a1, .LBB47_9
|
|
; RV32IFD-NEXT: .LBB47_6:
|
|
; RV32IFD-NEXT: seqz a2, a0
|
|
; RV32IFD-NEXT: j .LBB47_10
|
|
; RV32IFD-NEXT: .LBB47_7:
|
|
; RV32IFD-NEXT: srl a2, a2, a0
|
|
; RV32IFD-NEXT: not a0, a0
|
|
; RV32IFD-NEXT: slli a1, a1, 1
|
|
; RV32IFD-NEXT: sll a0, a1, a0
|
|
; RV32IFD-NEXT: or s4, a2, a0
|
|
; RV32IFD-NEXT: .LBB47_8: # %fp-to-i-if-exp.small
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s1
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s2, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s1
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s5, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s5, a0
|
|
; RV32IFD-NEXT: add s7, a1, a0
|
|
; RV32IFD-NEXT: mv a0, s4
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add s6, a0, s5
|
|
; RV32IFD-NEXT: sltu a0, s6, a0
|
|
; RV32IFD-NEXT: add a0, a1, a0
|
|
; RV32IFD-NEXT: add s8, s7, a0
|
|
; RV32IFD-NEXT: mv a0, s3
|
|
; RV32IFD-NEXT: li a1, 0
|
|
; RV32IFD-NEXT: mv a2, s0
|
|
; RV32IFD-NEXT: li a3, 0
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: mv s1, a0
|
|
; RV32IFD-NEXT: mv s5, a1
|
|
; RV32IFD-NEXT: add s9, a0, s8
|
|
; RV32IFD-NEXT: mv a0, s0
|
|
; RV32IFD-NEXT: mv a1, s0
|
|
; RV32IFD-NEXT: mv a2, s4
|
|
; RV32IFD-NEXT: mv a3, s3
|
|
; RV32IFD-NEXT: call __muldi3
|
|
; RV32IFD-NEXT: add a0, s9, a0
|
|
; RV32IFD-NEXT: sltu a2, s9, s1
|
|
; RV32IFD-NEXT: sltu a3, s8, s7
|
|
; RV32IFD-NEXT: sltu a4, a0, s9
|
|
; RV32IFD-NEXT: add a3, s5, a3
|
|
; RV32IFD-NEXT: add a2, a3, a2
|
|
; RV32IFD-NEXT: add a1, a2, a1
|
|
; RV32IFD-NEXT: add a1, a1, a4
|
|
; RV32IFD-NEXT: beqz a1, .LBB47_6
|
|
; RV32IFD-NEXT: .LBB47_9: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: srli a2, a1, 31
|
|
; RV32IFD-NEXT: .LBB47_10: # %fp-to-i-cleanup
|
|
; RV32IFD-NEXT: xori a0, a0, 1
|
|
; RV32IFD-NEXT: or a0, a0, a1
|
|
; RV32IFD-NEXT: seqz a0, a0
|
|
; RV32IFD-NEXT: addi a0, a0, -1
|
|
; RV32IFD-NEXT: and a0, a0, a2
|
|
; RV32IFD-NEXT: neg a0, a0
|
|
; RV32IFD-NEXT: and a2, a0, s6
|
|
; RV32IFD-NEXT: and a3, a0, s2
|
|
; RV32IFD-NEXT: and a0, a0, a1
|
|
; RV32IFD-NEXT: srli a0, a0, 31
|
|
; RV32IFD-NEXT: addi a1, a0, -1
|
|
; RV32IFD-NEXT: and a0, a1, a3
|
|
; RV32IFD-NEXT: and a1, a1, a2
|
|
; RV32IFD-NEXT: lw ra, 92(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s0, 88(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s1, 84(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s2, 80(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s3, 76(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s4, 72(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s5, 68(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s6, 64(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s7, 60(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s8, 56(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: lw s9, 52(sp) # 4-byte Folded Reload
|
|
; RV32IFD-NEXT: .cfi_restore ra
|
|
; RV32IFD-NEXT: .cfi_restore s0
|
|
; RV32IFD-NEXT: .cfi_restore s1
|
|
; RV32IFD-NEXT: .cfi_restore s2
|
|
; RV32IFD-NEXT: .cfi_restore s3
|
|
; RV32IFD-NEXT: .cfi_restore s4
|
|
; RV32IFD-NEXT: .cfi_restore s5
|
|
; RV32IFD-NEXT: .cfi_restore s6
|
|
; RV32IFD-NEXT: .cfi_restore s7
|
|
; RV32IFD-NEXT: .cfi_restore s8
|
|
; RV32IFD-NEXT: .cfi_restore s9
|
|
; RV32IFD-NEXT: addi sp, sp, 96
|
|
; RV32IFD-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32IFD-NEXT: ret
|
|
entry:
|
|
%conv = fptosi double %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 18446744073709551616)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 0)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @stest_f32i64_mm(float %x) {
|
|
; RV32-LABEL: stest_f32i64_mm:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: addi sp, sp, -80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: .cfi_offset s0, -8
|
|
; RV32-NEXT: .cfi_offset s1, -12
|
|
; RV32-NEXT: .cfi_offset s2, -16
|
|
; RV32-NEXT: .cfi_offset s3, -20
|
|
; RV32-NEXT: .cfi_offset s4, -24
|
|
; RV32-NEXT: .cfi_offset s5, -28
|
|
; RV32-NEXT: .cfi_offset s6, -32
|
|
; RV32-NEXT: .cfi_offset s7, -36
|
|
; RV32-NEXT: .cfi_offset s8, -40
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB48_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li s1, 0
|
|
; RV32-NEXT: li s2, 0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: j .LBB48_5
|
|
; RV32-NEXT: .LBB48_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: srai s0, a1, 31
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: or a1, a1, a2
|
|
; RV32-NEXT: ori s2, s0, 1
|
|
; RV32-NEXT: bltu a3, a0, .LBB48_4
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a2, 150
|
|
; RV32-NEXT: sub a2, a2, a0
|
|
; RV32-NEXT: srl s3, a1, a2
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s4
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add s6, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s7, s6, a0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s4, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: add s8, a0, s7
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s3
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a0, s8, a0
|
|
; RV32-NEXT: sltu a2, s8, s4
|
|
; RV32-NEXT: sltu a3, s7, s6
|
|
; RV32-NEXT: sltu a4, a0, s8
|
|
; RV32-NEXT: add a3, s5, a3
|
|
; RV32-NEXT: add a2, a3, a2
|
|
; RV32-NEXT: add a1, a2, a1
|
|
; RV32-NEXT: add a1, a1, a4
|
|
; RV32-NEXT: j .LBB48_5
|
|
; RV32-NEXT: .LBB48_4: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a1, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 32(sp)
|
|
; RV32-NEXT: sw zero, 36(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: sw zero, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: addi a1, sp, 24
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 0(a1)
|
|
; RV32-NEXT: lw a2, 4(a1)
|
|
; RV32-NEXT: lw a5, 8(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a6, a4, 1
|
|
; RV32-NEXT: srli a7, a2, 1
|
|
; RV32-NEXT: srli t0, a5, 1
|
|
; RV32-NEXT: srl a6, a6, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl a3, t0, a3
|
|
; RV32-NEXT: sll a2, a2, a0
|
|
; RV32-NEXT: sll a5, a5, a0
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or s3, a2, a6
|
|
; RV32-NEXT: or a2, a5, a7
|
|
; RV32-NEXT: or a3, a1, a3
|
|
; RV32-NEXT: sll s4, a4, a0
|
|
; RV32-NEXT: mv a0, s2
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s4
|
|
; RV32-NEXT: mv a3, s3
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s7, a0, s1
|
|
; RV32-NEXT: sltu a0, s7, a0
|
|
; RV32-NEXT: add s8, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s5, a0
|
|
; RV32-NEXT: mv s6, a1
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: add s2, s5, a1
|
|
; RV32-NEXT: sltu a0, s2, s5
|
|
; RV32-NEXT: add s6, s6, a0
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a0, s2, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s4, s6, a0
|
|
; RV32-NEXT: sltu s5, s4, s6
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s4, a0, s4
|
|
; RV32-NEXT: sltu a2, s4, a0
|
|
; RV32-NEXT: add a0, s4, s7
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: add a1, a1, s8
|
|
; RV32-NEXT: sltu a2, a0, s4
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: .LBB48_5: # %fp-to-i-cleanup
|
|
; RV32-NEXT: lui a2, 524288
|
|
; RV32-NEXT: addi a3, a2, -1
|
|
; RV32-NEXT: beq s2, a3, .LBB48_7
|
|
; RV32-NEXT: # %bb.6: # %fp-to-i-cleanup
|
|
; RV32-NEXT: sltu a4, s2, a3
|
|
; RV32-NEXT: or a5, a0, a1
|
|
; RV32-NEXT: bnez a5, .LBB48_8
|
|
; RV32-NEXT: j .LBB48_9
|
|
; RV32-NEXT: .LBB48_7:
|
|
; RV32-NEXT: sltiu a4, s1, -1
|
|
; RV32-NEXT: or a5, a0, a1
|
|
; RV32-NEXT: beqz a5, .LBB48_9
|
|
; RV32-NEXT: .LBB48_8: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a4, a1, 31
|
|
; RV32-NEXT: .LBB48_9: # %fp-to-i-cleanup
|
|
; RV32-NEXT: neg a5, a4
|
|
; RV32-NEXT: addi a6, a4, -1
|
|
; RV32-NEXT: bnez a4, .LBB48_11
|
|
; RV32-NEXT: # %bb.10: # %fp-to-i-cleanup
|
|
; RV32-NEXT: mv s2, a3
|
|
; RV32-NEXT: .LBB48_11: # %fp-to-i-cleanup
|
|
; RV32-NEXT: or a3, a6, s1
|
|
; RV32-NEXT: and a1, a5, a1
|
|
; RV32-NEXT: and a4, a5, a0
|
|
; RV32-NEXT: beq s2, a2, .LBB48_13
|
|
; RV32-NEXT: # %bb.12: # %fp-to-i-cleanup
|
|
; RV32-NEXT: sltu a0, a2, s2
|
|
; RV32-NEXT: j .LBB48_14
|
|
; RV32-NEXT: .LBB48_13:
|
|
; RV32-NEXT: snez a0, a3
|
|
; RV32-NEXT: .LBB48_14: # %fp-to-i-cleanup
|
|
; RV32-NEXT: and a4, a4, a1
|
|
; RV32-NEXT: li a2, -1
|
|
; RV32-NEXT: beq a4, a2, .LBB48_16
|
|
; RV32-NEXT: # %bb.15: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a1, a1, 31
|
|
; RV32-NEXT: xori a0, a1, 1
|
|
; RV32-NEXT: .LBB48_16: # %fp-to-i-cleanup
|
|
; RV32-NEXT: bnez a0, .LBB48_18
|
|
; RV32-NEXT: # %bb.17: # %fp-to-i-cleanup
|
|
; RV32-NEXT: lui s2, 524288
|
|
; RV32-NEXT: .LBB48_18: # %fp-to-i-cleanup
|
|
; RV32-NEXT: neg a0, a0
|
|
; RV32-NEXT: and a0, a0, a3
|
|
; RV32-NEXT: mv a1, s2
|
|
; RV32-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: .cfi_restore s0
|
|
; RV32-NEXT: .cfi_restore s1
|
|
; RV32-NEXT: .cfi_restore s2
|
|
; RV32-NEXT: .cfi_restore s3
|
|
; RV32-NEXT: .cfi_restore s4
|
|
; RV32-NEXT: .cfi_restore s5
|
|
; RV32-NEXT: .cfi_restore s6
|
|
; RV32-NEXT: .cfi_restore s7
|
|
; RV32-NEXT: .cfi_restore s8
|
|
; RV32-NEXT: addi sp, sp, 80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f32i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: feq.s a1, fa0, fa0
|
|
; RV64-NEXT: seqz a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 9223372036854775807)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 -9223372036854775808)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utest_f32i64_mm(float %x) {
|
|
; RV32-LABEL: utest_f32i64_mm:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB49_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: li a2, 0
|
|
; RV32-NEXT: j .LBB49_5
|
|
; RV32-NEXT: .LBB49_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: or a4, a1, a2
|
|
; RV32-NEXT: bltu a3, a0, .LBB49_4
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: li a2, 0
|
|
; RV32-NEXT: li a5, 150
|
|
; RV32-NEXT: sub a5, a5, a0
|
|
; RV32-NEXT: srl a0, a4, a5
|
|
; RV32-NEXT: j .LBB49_5
|
|
; RV32-NEXT: .LBB49_4: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi sp, sp, -32
|
|
; RV32-NEXT: .cfi_def_cfa_offset 32
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a4, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: sw zero, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 0(sp)
|
|
; RV32-NEXT: sw zero, 4(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: addi a1, sp, 16
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 8(a1)
|
|
; RV32-NEXT: lw a5, 4(a1)
|
|
; RV32-NEXT: lw a6, 0(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a2, a4, 1
|
|
; RV32-NEXT: srli a7, a5, 1
|
|
; RV32-NEXT: srli t0, a6, 1
|
|
; RV32-NEXT: srl a2, a2, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl t0, t0, a3
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or a2, a1, a2
|
|
; RV32-NEXT: sll a1, a4, a0
|
|
; RV32-NEXT: sll a4, a5, a0
|
|
; RV32-NEXT: or a3, a1, a7
|
|
; RV32-NEXT: or a1, a4, t0
|
|
; RV32-NEXT: sll a0, a6, a0
|
|
; RV32-NEXT: addi sp, sp, 32
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: .LBB49_5: # %fp-to-i-cleanup
|
|
; RV32-NEXT: or a4, a2, a3
|
|
; RV32-NEXT: xori a3, a3, 1
|
|
; RV32-NEXT: seqz a4, a4
|
|
; RV32-NEXT: or a2, a3, a2
|
|
; RV32-NEXT: seqz a2, a2
|
|
; RV32-NEXT: addi a2, a2, -1
|
|
; RV32-NEXT: and a2, a2, a4
|
|
; RV32-NEXT: neg a2, a2
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utest_f32i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixunssfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui float %x to i128
|
|
%spec.store.select = call i128 @llvm.umin.i128(i128 %conv, i128 18446744073709551616)
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f32i64_mm(float %x) {
|
|
; RV32-LABEL: ustest_f32i64_mm:
|
|
; RV32: # %bb.0: # %entryfp-to-i-entry
|
|
; RV32-NEXT: addi sp, sp, -80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 80
|
|
; RV32-NEXT: sw ra, 76(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s0, 72(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s1, 68(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s2, 64(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s3, 60(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s4, 56(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s5, 52(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s6, 48(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s7, 44(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: sw s8, 40(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: .cfi_offset s0, -8
|
|
; RV32-NEXT: .cfi_offset s1, -12
|
|
; RV32-NEXT: .cfi_offset s2, -16
|
|
; RV32-NEXT: .cfi_offset s3, -20
|
|
; RV32-NEXT: .cfi_offset s4, -24
|
|
; RV32-NEXT: .cfi_offset s5, -28
|
|
; RV32-NEXT: .cfi_offset s6, -32
|
|
; RV32-NEXT: .cfi_offset s7, -36
|
|
; RV32-NEXT: .cfi_offset s8, -40
|
|
; RV32-NEXT: fmv.x.w a1, fa0
|
|
; RV32-NEXT: slli a0, a1, 1
|
|
; RV32-NEXT: srli a0, a0, 24
|
|
; RV32-NEXT: li a2, 127
|
|
; RV32-NEXT: bgeu a0, a2, .LBB50_2
|
|
; RV32-NEXT: # %bb.1:
|
|
; RV32-NEXT: li s1, 0
|
|
; RV32-NEXT: li s5, 0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: j .LBB50_4
|
|
; RV32-NEXT: .LBB50_2: # %fp-to-i-if-check.exp.size
|
|
; RV32-NEXT: srai s0, a1, 31
|
|
; RV32-NEXT: slli a1, a1, 9
|
|
; RV32-NEXT: lui a2, 2048
|
|
; RV32-NEXT: li a3, 149
|
|
; RV32-NEXT: srli a1, a1, 9
|
|
; RV32-NEXT: or a1, a1, a2
|
|
; RV32-NEXT: ori s2, s0, 1
|
|
; RV32-NEXT: bltu a3, a0, .LBB50_5
|
|
; RV32-NEXT: # %bb.3: # %fp-to-i-if-exp.small
|
|
; RV32-NEXT: li a2, 150
|
|
; RV32-NEXT: sub a2, a2, a0
|
|
; RV32-NEXT: srl s3, a1, a2
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s4, a0, s4
|
|
; RV32-NEXT: sltu a0, s4, a0
|
|
; RV32-NEXT: add s6, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s5, a0, s4
|
|
; RV32-NEXT: sltu a0, s5, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s7, s6, a0
|
|
; RV32-NEXT: li a0, 0
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s2, a0
|
|
; RV32-NEXT: mv s4, a1
|
|
; RV32-NEXT: add s8, a0, s7
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s3
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a0, s8, a0
|
|
; RV32-NEXT: sltu a2, s8, s2
|
|
; RV32-NEXT: sltu a3, s7, s6
|
|
; RV32-NEXT: sltu a4, a0, s8
|
|
; RV32-NEXT: add a3, s4, a3
|
|
; RV32-NEXT: add a2, a3, a2
|
|
; RV32-NEXT: add a1, a2, a1
|
|
; RV32-NEXT: add a1, a1, a4
|
|
; RV32-NEXT: bnez a1, .LBB50_6
|
|
; RV32-NEXT: .LBB50_4:
|
|
; RV32-NEXT: seqz a2, a0
|
|
; RV32-NEXT: j .LBB50_7
|
|
; RV32-NEXT: .LBB50_5: # %fp-to-i-if-exp.large
|
|
; RV32-NEXT: addi a0, a0, -150
|
|
; RV32-NEXT: sw a1, 24(sp)
|
|
; RV32-NEXT: sw zero, 28(sp)
|
|
; RV32-NEXT: sw zero, 32(sp)
|
|
; RV32-NEXT: sw zero, 36(sp)
|
|
; RV32-NEXT: sw zero, 8(sp)
|
|
; RV32-NEXT: sw zero, 12(sp)
|
|
; RV32-NEXT: sw zero, 16(sp)
|
|
; RV32-NEXT: sw zero, 20(sp)
|
|
; RV32-NEXT: addi a1, sp, 24
|
|
; RV32-NEXT: srli a2, a0, 3
|
|
; RV32-NEXT: andi a3, a0, 31
|
|
; RV32-NEXT: andi a2, a2, 12
|
|
; RV32-NEXT: xori a3, a3, 31
|
|
; RV32-NEXT: sub a1, a1, a2
|
|
; RV32-NEXT: lw a4, 0(a1)
|
|
; RV32-NEXT: lw a2, 4(a1)
|
|
; RV32-NEXT: lw a5, 8(a1)
|
|
; RV32-NEXT: lw a1, 12(a1)
|
|
; RV32-NEXT: srli a6, a4, 1
|
|
; RV32-NEXT: srli a7, a2, 1
|
|
; RV32-NEXT: srli t0, a5, 1
|
|
; RV32-NEXT: srl a6, a6, a3
|
|
; RV32-NEXT: srl a7, a7, a3
|
|
; RV32-NEXT: srl a3, t0, a3
|
|
; RV32-NEXT: sll a2, a2, a0
|
|
; RV32-NEXT: sll a5, a5, a0
|
|
; RV32-NEXT: sll a1, a1, a0
|
|
; RV32-NEXT: or s3, a2, a6
|
|
; RV32-NEXT: or a2, a5, a7
|
|
; RV32-NEXT: or a3, a1, a3
|
|
; RV32-NEXT: sll s4, a4, a0
|
|
; RV32-NEXT: mv a0, s2
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: mv s5, a1
|
|
; RV32-NEXT: mv a0, s0
|
|
; RV32-NEXT: mv a1, s0
|
|
; RV32-NEXT: mv a2, s4
|
|
; RV32-NEXT: mv a3, s3
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s5
|
|
; RV32-NEXT: add s7, a0, s1
|
|
; RV32-NEXT: sltu a0, s7, a0
|
|
; RV32-NEXT: add s8, a1, a0
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s5, a0
|
|
; RV32-NEXT: mv s6, a1
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s2
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: mv s1, a0
|
|
; RV32-NEXT: add s2, s5, a1
|
|
; RV32-NEXT: sltu a0, s2, s5
|
|
; RV32-NEXT: add s6, s6, a0
|
|
; RV32-NEXT: mv a0, s4
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add s5, a0, s2
|
|
; RV32-NEXT: sltu a0, s5, a0
|
|
; RV32-NEXT: add a0, a1, a0
|
|
; RV32-NEXT: add s2, s6, a0
|
|
; RV32-NEXT: sltu s4, s2, s6
|
|
; RV32-NEXT: mv a0, s3
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: mv a2, s0
|
|
; RV32-NEXT: li a3, 0
|
|
; RV32-NEXT: call __muldi3
|
|
; RV32-NEXT: add a1, a1, s4
|
|
; RV32-NEXT: add s2, a0, s2
|
|
; RV32-NEXT: sltu a2, s2, a0
|
|
; RV32-NEXT: add a0, s2, s7
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: add a1, a1, s8
|
|
; RV32-NEXT: sltu a2, a0, s2
|
|
; RV32-NEXT: add a1, a1, a2
|
|
; RV32-NEXT: beqz a1, .LBB50_4
|
|
; RV32-NEXT: .LBB50_6: # %fp-to-i-cleanup
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: .LBB50_7: # %fp-to-i-cleanup
|
|
; RV32-NEXT: xori a0, a0, 1
|
|
; RV32-NEXT: or a0, a0, a1
|
|
; RV32-NEXT: seqz a0, a0
|
|
; RV32-NEXT: addi a0, a0, -1
|
|
; RV32-NEXT: and a0, a0, a2
|
|
; RV32-NEXT: neg a0, a0
|
|
; RV32-NEXT: and a2, a0, s5
|
|
; RV32-NEXT: and s1, a0, s1
|
|
; RV32-NEXT: and a0, a0, a1
|
|
; RV32-NEXT: srli a0, a0, 31
|
|
; RV32-NEXT: addi a1, a0, -1
|
|
; RV32-NEXT: and a0, a1, s1
|
|
; RV32-NEXT: and a1, a1, a2
|
|
; RV32-NEXT: lw ra, 76(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s0, 72(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s1, 68(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s2, 64(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s3, 60(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s4, 56(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s5, 52(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s6, 48(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s7, 44(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: lw s8, 40(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: .cfi_restore s0
|
|
; RV32-NEXT: .cfi_restore s1
|
|
; RV32-NEXT: .cfi_restore s2
|
|
; RV32-NEXT: .cfi_restore s3
|
|
; RV32-NEXT: .cfi_restore s4
|
|
; RV32-NEXT: .cfi_restore s5
|
|
; RV32-NEXT: .cfi_restore s6
|
|
; RV32-NEXT: .cfi_restore s7
|
|
; RV32-NEXT: .cfi_restore s8
|
|
; RV32-NEXT: addi sp, sp, 80
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f32i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: mv a2, a1
|
|
; RV64-NEXT: blez a1, .LBB50_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a2, 1
|
|
; RV64-NEXT: .LBB50_2: # %entry
|
|
; RV64-NEXT: slti a1, a1, 1
|
|
; RV64-NEXT: srli a2, a2, 63
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: addi a2, a2, -1
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi float %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 18446744073709551616)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 0)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @stest_f16i64_mm(half %x) {
|
|
; RV32-LABEL: stest_f16i64_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: srai a1, a0, 31
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: stest_f16i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: li a2, -1
|
|
; RV64-NEXT: srli a3, a2, 1
|
|
; RV64-NEXT: beqz a1, .LBB51_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: srli a4, a1, 63
|
|
; RV64-NEXT: j .LBB51_3
|
|
; RV64-NEXT: .LBB51_2:
|
|
; RV64-NEXT: sltu a4, a0, a3
|
|
; RV64-NEXT: .LBB51_3: # %entry
|
|
; RV64-NEXT: neg a5, a4
|
|
; RV64-NEXT: and a5, a5, a1
|
|
; RV64-NEXT: bnez a4, .LBB51_5
|
|
; RV64-NEXT: # %bb.4: # %entry
|
|
; RV64-NEXT: mv a0, a3
|
|
; RV64-NEXT: .LBB51_5: # %entry
|
|
; RV64-NEXT: slli a1, a2, 63
|
|
; RV64-NEXT: beq a5, a2, .LBB51_7
|
|
; RV64-NEXT: # %bb.6: # %entry
|
|
; RV64-NEXT: srli a5, a5, 63
|
|
; RV64-NEXT: xori a2, a5, 1
|
|
; RV64-NEXT: beqz a2, .LBB51_8
|
|
; RV64-NEXT: j .LBB51_9
|
|
; RV64-NEXT: .LBB51_7:
|
|
; RV64-NEXT: sltu a2, a1, a0
|
|
; RV64-NEXT: bnez a2, .LBB51_9
|
|
; RV64-NEXT: .LBB51_8: # %entry
|
|
; RV64-NEXT: mv a0, a1
|
|
; RV64-NEXT: .LBB51_9: # %entry
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 9223372036854775807)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 -9223372036854775808)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @utesth_f16i64_mm(half %x) {
|
|
; RV32-LABEL: utesth_f16i64_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.wu.s a0, fa0, rtz
|
|
; RV32-NEXT: li a1, 0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: utesth_f16i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixunssfti
|
|
; RV64-NEXT: snez a1, a1
|
|
; RV64-NEXT: addi a1, a1, -1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptoui half %x to i128
|
|
%spec.store.select = call i128 @llvm.umin.i128(i128 %conv, i128 18446744073709551616)
|
|
%conv6 = trunc i128 %spec.store.select to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
define i64 @ustest_f16i64_mm(half %x) {
|
|
; RV32-LABEL: ustest_f16i64_mm:
|
|
; RV32: # %bb.0: # %entry
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: srai a1, a0, 31
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: seqz a3, a1
|
|
; RV32-NEXT: or a2, a3, a2
|
|
; RV32-NEXT: ori a3, a1, 1
|
|
; RV32-NEXT: seqz a3, a3
|
|
; RV32-NEXT: addi a3, a3, -1
|
|
; RV32-NEXT: and a2, a3, a2
|
|
; RV32-NEXT: neg a2, a2
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: srli a2, a1, 31
|
|
; RV32-NEXT: addi a2, a2, -1
|
|
; RV32-NEXT: and a0, a2, a0
|
|
; RV32-NEXT: and a1, a2, a1
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i64_mm:
|
|
; RV64: # %bb.0: # %entry
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: call __fixsfti
|
|
; RV64-NEXT: mv a2, a1
|
|
; RV64-NEXT: blez a1, .LBB53_2
|
|
; RV64-NEXT: # %bb.1: # %entry
|
|
; RV64-NEXT: li a2, 1
|
|
; RV64-NEXT: .LBB53_2: # %entry
|
|
; RV64-NEXT: slti a1, a1, 1
|
|
; RV64-NEXT: srli a2, a2, 63
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: addi a2, a2, -1
|
|
; RV64-NEXT: and a0, a2, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
entry:
|
|
%conv = fptosi half %x to i128
|
|
%spec.store.select = call i128 @llvm.smin.i128(i128 %conv, i128 18446744073709551616)
|
|
%spec.store.select7 = call i128 @llvm.smax.i128(i128 %spec.store.select, i128 0)
|
|
%conv6 = trunc i128 %spec.store.select7 to i64
|
|
ret i64 %conv6
|
|
}
|
|
|
|
; i32 non saturate
|
|
|
|
define i32 @ustest_f16i32_nsat(half %x) {
|
|
; RV32-LABEL: ustest_f16i32_nsat:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: addi sp, sp, -16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 16
|
|
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32-NEXT: .cfi_offset ra, -4
|
|
; RV32-NEXT: call __extendhfsf2
|
|
; RV32-NEXT: fcvt.w.s a0, fa0, rtz
|
|
; RV32-NEXT: srai a1, a0, 31
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: sgtz a1, a0
|
|
; RV32-NEXT: neg a1, a1
|
|
; RV32-NEXT: and a0, a1, a0
|
|
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32-NEXT: .cfi_restore ra
|
|
; RV32-NEXT: addi sp, sp, 16
|
|
; RV32-NEXT: .cfi_def_cfa_offset 0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-LABEL: ustest_f16i32_nsat:
|
|
; RV64: # %bb.0:
|
|
; RV64-NEXT: addi sp, sp, -16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 16
|
|
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64-NEXT: .cfi_offset ra, -8
|
|
; RV64-NEXT: call __extendhfsf2
|
|
; RV64-NEXT: fcvt.l.s a0, fa0, rtz
|
|
; RV64-NEXT: srai a1, a0, 63
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: sgtz a1, a0
|
|
; RV64-NEXT: neg a1, a1
|
|
; RV64-NEXT: and a0, a1, a0
|
|
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64-NEXT: .cfi_restore ra
|
|
; RV64-NEXT: addi sp, sp, 16
|
|
; RV64-NEXT: .cfi_def_cfa_offset 0
|
|
; RV64-NEXT: ret
|
|
%conv = fptosi half %x to i32
|
|
%spec.store.select = call i32 @llvm.smin.i32(i32 0, i32 %conv)
|
|
%spec.store.select7 = call i32 @llvm.smax.i32(i32 %spec.store.select, i32 0)
|
|
ret i32 %spec.store.select7
|
|
}
|
|
|