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
548 lines
17 KiB
LLVM
548 lines
17 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+f,+d,+zvfhmin,+zvfbfmin \
|
|
; RUN: -target-abi=ilp32d -verify-machineinstrs | FileCheck %s --check-prefix=RV32
|
|
; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv64 -mattr=+v,+f,+d,+zvfhmin,+zvfbfmin \
|
|
; RUN: -target-abi=lp64d -verify-machineinstrs | FileCheck %s --check-prefix=RV64-i32
|
|
; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+f,+d,+zvfhmin,+zvfbfmin \
|
|
; RUN: -target-abi=lp64d -verify-machineinstrs | FileCheck %s --check-prefix=RV64-i64
|
|
|
|
define <1 x iXLen> @lround_v1f16(<1 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v1f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 1, e16, mf4, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v1f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 1, e16, mf4, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v1f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 1, e16, mf4, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v9
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <1 x iXLen> @llvm.lround.v1iXLen.v1f16(<1 x half> %x)
|
|
ret <1 x iXLen> %a
|
|
}
|
|
|
|
define <2 x iXLen> @lround_v2f16(<2 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v2f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v2f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v2f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v9
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <2 x iXLen> @llvm.lround.v2iXLen.v2f16(<2 x half> %x)
|
|
ret <2 x iXLen> %a
|
|
}
|
|
|
|
define <3 x iXLen> @lround_v3f16(<3 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v3f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v3f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v3f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v10, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v10
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <3 x iXLen> @llvm.lround.v3iXLen.v3f16(<3 x half> %x)
|
|
ret <3 x iXLen> %a
|
|
}
|
|
|
|
define <4 x iXLen> @lround_v4f16(<4 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v4f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v4f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v9, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v9
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v4f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v10, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v10
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <4 x iXLen> @llvm.lround.v4iXLen.v4f16(<4 x half> %x)
|
|
ret <4 x iXLen> %a
|
|
}
|
|
|
|
define <8 x iXLen> @lround_v8f16(<8 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v8f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 8, e16, m1, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v10, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v10
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v8f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 8, e16, m1, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v10, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v10
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v8f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 8, e16, m1, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v12, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v12
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <8 x iXLen> @llvm.lround.v8iXLen.v8f16(<8 x half> %x)
|
|
ret <8 x iXLen> %a
|
|
}
|
|
|
|
define <16 x iXLen> @lround_v16f16(<16 x half> %x) nounwind {
|
|
; RV32-LABEL: lround_v16f16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: vsetivli zero, 16, e16, m2, ta, ma
|
|
; RV32-NEXT: vfwcvt.f.f.v v12, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, m4, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v12
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v16f16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: vsetivli zero, 16, e16, m2, ta, ma
|
|
; RV64-i32-NEXT: vfwcvt.f.f.v v12, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, m4, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v12
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v16f16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 16, e16, m2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.f.f.v v16, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, m4, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v16
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <16 x iXLen> @llvm.lround.v16iXLen.v16f16(<16 x half> %x)
|
|
ret <16 x iXLen> %a
|
|
}
|
|
|
|
define <1 x iXLen> @lround_v1f32(<1 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v1f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v1f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v1f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v9, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: vmv1r.v v8, v9
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <1 x iXLen> @llvm.lround.v1iXLen.v1f32(<1 x float> %x)
|
|
ret <1 x iXLen> %a
|
|
}
|
|
|
|
define <2 x iXLen> @lround_v2f32(<2 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v2f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v2f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v2f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v9, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: vmv1r.v v8, v9
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <2 x iXLen> @llvm.lround.v2iXLen.v2f32(<2 x float> %x)
|
|
ret <2 x iXLen> %a
|
|
}
|
|
|
|
define <3 x iXLen> @lround_v3f32(<3 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v3f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v3f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v3f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV64-i64-NEXT: vmv1r.v v10, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v10
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <3 x iXLen> @llvm.lround.v3iXLen.v3f32(<3 x float> %x)
|
|
ret <3 x iXLen> %a
|
|
}
|
|
|
|
define <4 x iXLen> @lround_v4f32(<4 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v4f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v4f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v4f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV64-i64-NEXT: vmv1r.v v10, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v10
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <4 x iXLen> @llvm.lround.v4iXLen.v4f32(<4 x float> %x)
|
|
ret <4 x iXLen> %a
|
|
}
|
|
|
|
define <8 x iXLen> @lround_v8f32(<8 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v8f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v8f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 8, e32, m2, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v8f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 8, e32, m2, ta, ma
|
|
; RV64-i64-NEXT: vmv2r.v v12, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v12
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <8 x iXLen> @llvm.lround.v8iXLen.v8f32(<8 x float> %x)
|
|
ret <8 x iXLen> %a
|
|
}
|
|
|
|
define <16 x iXLen> @lround_v16f32(<16 x float> %x) nounwind {
|
|
; RV32-LABEL: lround_v16f32:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 16, e32, m4, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v16f32:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 16, e32, m4, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v16f32:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 16, e32, m4, ta, ma
|
|
; RV64-i64-NEXT: vmv4r.v v16, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v16
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <16 x iXLen> @llvm.lround.v16iXLen.v16f32(<16 x float> %x)
|
|
ret <16 x iXLen> %a
|
|
}
|
|
|
|
define <1 x iXLen> @lround_v1f64(<1 x double> %x) nounwind {
|
|
; RV32-LABEL: lround_v1f64:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfncvt.x.f.w v9, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: vmv1r.v v8, v9
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v1f64:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfncvt.x.f.w v9, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: vmv1r.v v8, v9
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v1f64:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 1, e64, m1, ta, ma
|
|
; RV64-i64-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <1 x iXLen> @llvm.lround.v1iXLen.v1f64(<1 x double> %x)
|
|
ret <1 x iXLen> %a
|
|
}
|
|
|
|
define <2 x iXLen> @lround_v2f64(<2 x double> %x) nounwind {
|
|
; RV32-LABEL: lround_v2f64:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
|
|
; RV32-NEXT: vfncvt.x.f.w v9, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: vmv1r.v v8, v9
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v2f64:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
|
|
; RV64-i32-NEXT: vfncvt.x.f.w v9, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: vmv1r.v v8, v9
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v2f64:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 2, e64, m1, ta, ma
|
|
; RV64-i64-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <2 x iXLen> @llvm.lround.v2iXLen.v2f64(<2 x double> %x)
|
|
ret <2 x iXLen> %a
|
|
}
|
|
|
|
define <4 x iXLen> @lround_v4f64(<4 x double> %x) nounwind {
|
|
; RV32-LABEL: lround_v4f64:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV32-NEXT: vfncvt.x.f.w v10, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: vmv.v.v v8, v10
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v4f64:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 4, e32, m1, ta, ma
|
|
; RV64-i32-NEXT: vfncvt.x.f.w v10, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: vmv.v.v v8, v10
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v4f64:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 4, e64, m2, ta, ma
|
|
; RV64-i64-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <4 x iXLen> @llvm.lround.v4iXLen.v4f64(<4 x double> %x)
|
|
ret <4 x iXLen> %a
|
|
}
|
|
|
|
define <8 x iXLen> @lround_v8f64(<8 x double> %x) nounwind {
|
|
; RV32-LABEL: lround_v8f64:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetivli zero, 8, e32, m2, ta, ma
|
|
; RV32-NEXT: vfncvt.x.f.w v12, v8
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: vmv.v.v v8, v12
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v8f64:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetivli zero, 8, e32, m2, ta, ma
|
|
; RV64-i32-NEXT: vfncvt.x.f.w v12, v8
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: vmv.v.v v8, v12
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v8f64:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 8, e64, m4, ta, ma
|
|
; RV64-i64-NEXT: vfcvt.x.f.v v8, v8
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <8 x iXLen> @llvm.lround.v8iXLen.v8f64(<8 x double> %x)
|
|
ret <8 x iXLen> %a
|
|
}
|
|
|
|
define <32 x iXLen> @lround_v32bf16(<32 x bfloat> %x) {
|
|
; RV32-LABEL: lround_v32bf16:
|
|
; RV32: # %bb.0:
|
|
; RV32-NEXT: li a0, 32
|
|
; RV32-NEXT: vsetvli zero, a0, e16, m4, ta, ma
|
|
; RV32-NEXT: vfwcvtbf16.f.f.v v16, v8
|
|
; RV32-NEXT: fsrmi a0, 4
|
|
; RV32-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; RV32-NEXT: vfcvt.x.f.v v8, v16
|
|
; RV32-NEXT: fsrm a0
|
|
; RV32-NEXT: ret
|
|
;
|
|
; RV64-i32-LABEL: lround_v32bf16:
|
|
; RV64-i32: # %bb.0:
|
|
; RV64-i32-NEXT: li a0, 32
|
|
; RV64-i32-NEXT: vsetvli zero, a0, e16, m4, ta, ma
|
|
; RV64-i32-NEXT: vfwcvtbf16.f.f.v v16, v8
|
|
; RV64-i32-NEXT: fsrmi a0, 4
|
|
; RV64-i32-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; RV64-i32-NEXT: vfcvt.x.f.v v8, v16
|
|
; RV64-i32-NEXT: fsrm a0
|
|
; RV64-i32-NEXT: ret
|
|
;
|
|
; RV64-i64-LABEL: lround_v32bf16:
|
|
; RV64-i64: # %bb.0:
|
|
; RV64-i64-NEXT: vsetivli zero, 16, e16, m2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvtbf16.f.f.v v16, v8
|
|
; RV64-i64-NEXT: fsrmi a0, 4
|
|
; RV64-i64-NEXT: vsetivli zero, 16, e16, m4, ta, ma
|
|
; RV64-i64-NEXT: vslidedown.vi v20, v8, 16
|
|
; RV64-i64-NEXT: vsetivli zero, 16, e32, m4, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v8, v16
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e16, m2, ta, ma
|
|
; RV64-i64-NEXT: vfwcvtbf16.f.f.v v24, v20
|
|
; RV64-i64-NEXT: vsetvli zero, zero, e32, m4, ta, ma
|
|
; RV64-i64-NEXT: vfwcvt.x.f.v v16, v24
|
|
; RV64-i64-NEXT: fsrm a0
|
|
; RV64-i64-NEXT: ret
|
|
%a = call <32 x iXLen> @llvm.lround.v32iXLen.v32bf16(<32 x bfloat> %x)
|
|
ret <32 x iXLen> %a
|
|
}
|