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
274 lines
8.1 KiB
LLVM
274 lines
8.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=CHECK,RV32I %s
|
|
; RUN: llc -mtriple=riscv32 -mattr=+m -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=CHECK,RV32IM %s
|
|
; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=CHECK,RV64I %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+m -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=CHECK,RV64IM %s
|
|
|
|
define i32 @fold_urem_positive_odd(i32 %x) nounwind {
|
|
; RV32I-LABEL: fold_urem_positive_odd:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: li a1, 95
|
|
; RV32I-NEXT: tail __umodsi3
|
|
;
|
|
; RV32IM-LABEL: fold_urem_positive_odd:
|
|
; RV32IM: # %bb.0:
|
|
; RV32IM-NEXT: lui a1, 364242
|
|
; RV32IM-NEXT: addi a1, a1, 777
|
|
; RV32IM-NEXT: mulhu a1, a0, a1
|
|
; RV32IM-NEXT: sub a2, a0, a1
|
|
; RV32IM-NEXT: srli a2, a2, 1
|
|
; RV32IM-NEXT: add a1, a2, a1
|
|
; RV32IM-NEXT: srli a1, a1, 6
|
|
; RV32IM-NEXT: li a2, 95
|
|
; RV32IM-NEXT: mul a1, a1, a2
|
|
; RV32IM-NEXT: sub a0, a0, a1
|
|
; RV32IM-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: fold_urem_positive_odd:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: slli a0, a0, 32
|
|
; RV64I-NEXT: srli a0, a0, 32
|
|
; RV64I-NEXT: li a1, 95
|
|
; RV64I-NEXT: call __umoddi3
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IM-LABEL: fold_urem_positive_odd:
|
|
; RV64IM: # %bb.0:
|
|
; RV64IM-NEXT: slli a1, a0, 32
|
|
; RV64IM-NEXT: lui a2, 172
|
|
; RV64IM-NEXT: srli a1, a1, 32
|
|
; RV64IM-NEXT: addi a2, a2, 1897
|
|
; RV64IM-NEXT: slli a2, a2, 13
|
|
; RV64IM-NEXT: addi a2, a2, 777
|
|
; RV64IM-NEXT: slli a2, a2, 25
|
|
; RV64IM-NEXT: mulhu a1, a1, a2
|
|
; RV64IM-NEXT: li a2, 95
|
|
; RV64IM-NEXT: mul a1, a1, a2
|
|
; RV64IM-NEXT: subw a0, a0, a1
|
|
; RV64IM-NEXT: ret
|
|
%1 = urem i32 %x, 95
|
|
ret i32 %1
|
|
}
|
|
|
|
|
|
define i32 @fold_urem_positive_even(i32 %x) nounwind {
|
|
; RV32I-LABEL: fold_urem_positive_even:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: li a1, 1060
|
|
; RV32I-NEXT: tail __umodsi3
|
|
;
|
|
; RV32IM-LABEL: fold_urem_positive_even:
|
|
; RV32IM: # %bb.0:
|
|
; RV32IM-NEXT: lui a1, 1012964
|
|
; RV32IM-NEXT: addi a1, a1, -61
|
|
; RV32IM-NEXT: mulhu a1, a0, a1
|
|
; RV32IM-NEXT: srli a1, a1, 10
|
|
; RV32IM-NEXT: li a2, 1060
|
|
; RV32IM-NEXT: mul a1, a1, a2
|
|
; RV32IM-NEXT: sub a0, a0, a1
|
|
; RV32IM-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: fold_urem_positive_even:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: slli a0, a0, 32
|
|
; RV64I-NEXT: srli a0, a0, 32
|
|
; RV64I-NEXT: li a1, 1060
|
|
; RV64I-NEXT: call __umoddi3
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IM-LABEL: fold_urem_positive_even:
|
|
; RV64IM: # %bb.0:
|
|
; RV64IM-NEXT: slli a1, a0, 32
|
|
; RV64IM-NEXT: lui a2, 1012964
|
|
; RV64IM-NEXT: addi a2, a2, -61
|
|
; RV64IM-NEXT: slli a2, a2, 32
|
|
; RV64IM-NEXT: mulhu a1, a1, a2
|
|
; RV64IM-NEXT: srli a1, a1, 42
|
|
; RV64IM-NEXT: li a2, 1060
|
|
; RV64IM-NEXT: mul a1, a1, a2
|
|
; RV64IM-NEXT: subw a0, a0, a1
|
|
; RV64IM-NEXT: ret
|
|
%1 = urem i32 %x, 1060
|
|
ret i32 %1
|
|
}
|
|
|
|
|
|
; Don't fold if we can combine urem with udiv.
|
|
define i32 @combine_urem_udiv(i32 %x) nounwind {
|
|
; RV32I-LABEL: combine_urem_udiv:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: li a1, 95
|
|
; RV32I-NEXT: call __umodsi3
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: li a1, 95
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __udivsi3
|
|
; RV32I-NEXT: add a0, s1, a0
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IM-LABEL: combine_urem_udiv:
|
|
; RV32IM: # %bb.0:
|
|
; RV32IM-NEXT: lui a1, 364242
|
|
; RV32IM-NEXT: addi a1, a1, 777
|
|
; RV32IM-NEXT: mulhu a1, a0, a1
|
|
; RV32IM-NEXT: sub a2, a0, a1
|
|
; RV32IM-NEXT: srli a2, a2, 1
|
|
; RV32IM-NEXT: add a1, a2, a1
|
|
; RV32IM-NEXT: li a2, 95
|
|
; RV32IM-NEXT: srli a1, a1, 6
|
|
; RV32IM-NEXT: mul a2, a1, a2
|
|
; RV32IM-NEXT: add a0, a0, a1
|
|
; RV32IM-NEXT: sub a0, a0, a2
|
|
; RV32IM-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: combine_urem_udiv:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: slli a0, a0, 32
|
|
; RV64I-NEXT: srli s0, a0, 32
|
|
; RV64I-NEXT: li a1, 95
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __umoddi3
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: li a1, 95
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __udivdi3
|
|
; RV64I-NEXT: add a0, s1, a0
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IM-LABEL: combine_urem_udiv:
|
|
; RV64IM: # %bb.0:
|
|
; RV64IM-NEXT: slli a1, a0, 32
|
|
; RV64IM-NEXT: lui a2, 172
|
|
; RV64IM-NEXT: srli a1, a1, 32
|
|
; RV64IM-NEXT: addi a2, a2, 1897
|
|
; RV64IM-NEXT: slli a2, a2, 13
|
|
; RV64IM-NEXT: addi a2, a2, 777
|
|
; RV64IM-NEXT: slli a2, a2, 25
|
|
; RV64IM-NEXT: mulhu a1, a1, a2
|
|
; RV64IM-NEXT: li a2, 95
|
|
; RV64IM-NEXT: mul a2, a1, a2
|
|
; RV64IM-NEXT: add a0, a0, a1
|
|
; RV64IM-NEXT: subw a0, a0, a2
|
|
; RV64IM-NEXT: ret
|
|
%1 = urem i32 %x, 95
|
|
%2 = udiv i32 %x, 95
|
|
%3 = add i32 %1, %2
|
|
ret i32 %3
|
|
}
|
|
|
|
; Don't fold for divisors that are a power of two.
|
|
define i32 @dont_fold_urem_power_of_two(i32 %x) nounwind {
|
|
; CHECK-LABEL: dont_fold_urem_power_of_two:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: andi a0, a0, 63
|
|
; CHECK-NEXT: ret
|
|
%1 = urem i32 %x, 64
|
|
ret i32 %1
|
|
}
|
|
|
|
; Don't fold if the divisor is one.
|
|
define i32 @dont_fold_urem_one(i32 %x) nounwind {
|
|
; CHECK-LABEL: dont_fold_urem_one:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: li a0, 0
|
|
; CHECK-NEXT: ret
|
|
%1 = urem i32 %x, 1
|
|
ret i32 %1
|
|
}
|
|
|
|
; Don't fold if the divisor is 2^32.
|
|
define i32 @dont_fold_urem_i32_umax(i32 %x) nounwind {
|
|
; CHECK-LABEL: dont_fold_urem_i32_umax:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: ret
|
|
%1 = urem i32 %x, 4294967296
|
|
ret i32 %1
|
|
}
|
|
|
|
; Don't fold i64 urem
|
|
define i64 @dont_fold_urem_i64(i64 %x) nounwind {
|
|
; RV32I-LABEL: dont_fold_urem_i64:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 98
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __umoddi3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IM-LABEL: dont_fold_urem_i64:
|
|
; RV32IM: # %bb.0:
|
|
; RV32IM-NEXT: slli a2, a1, 10
|
|
; RV32IM-NEXT: srli a3, a0, 22
|
|
; RV32IM-NEXT: or a2, a3, a2
|
|
; RV32IM-NEXT: lui a3, 512
|
|
; RV32IM-NEXT: addi a3, a3, -1
|
|
; RV32IM-NEXT: and a2, a2, a3
|
|
; RV32IM-NEXT: slli a3, a0, 10
|
|
; RV32IM-NEXT: srli a1, a1, 11
|
|
; RV32IM-NEXT: srli a3, a3, 11
|
|
; RV32IM-NEXT: add a1, a3, a1
|
|
; RV32IM-NEXT: lui a3, 21400
|
|
; RV32IM-NEXT: add a1, a1, a2
|
|
; RV32IM-NEXT: li a2, 49
|
|
; RV32IM-NEXT: addi a3, a3, -2006
|
|
; RV32IM-NEXT: mulhu a3, a1, a3
|
|
; RV32IM-NEXT: mul a2, a3, a2
|
|
; RV32IM-NEXT: sub a1, a1, a2
|
|
; RV32IM-NEXT: slli a1, a1, 1
|
|
; RV32IM-NEXT: andi a0, a0, 1
|
|
; RV32IM-NEXT: or a0, a1, a0
|
|
; RV32IM-NEXT: li a1, 0
|
|
; RV32IM-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: dont_fold_urem_i64:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: li a1, 98
|
|
; RV64I-NEXT: tail __umoddi3
|
|
;
|
|
; RV64IM-LABEL: dont_fold_urem_i64:
|
|
; RV64IM: # %bb.0:
|
|
; RV64IM-NEXT: lui a1, %hi(.LCPI6_0)
|
|
; RV64IM-NEXT: ld a1, %lo(.LCPI6_0)(a1)
|
|
; RV64IM-NEXT: srli a2, a0, 1
|
|
; RV64IM-NEXT: mulhu a1, a2, a1
|
|
; RV64IM-NEXT: srli a1, a1, 4
|
|
; RV64IM-NEXT: li a2, 98
|
|
; RV64IM-NEXT: mul a1, a1, a2
|
|
; RV64IM-NEXT: sub a0, a0, a1
|
|
; RV64IM-NEXT: ret
|
|
%1 = urem i64 %x, 98
|
|
ret i64 %1
|
|
}
|