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
66 lines
2.4 KiB
LLVM
66 lines
2.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc < %s -mcpu=sm_90 -mattr=+ptx78 | FileCheck --check-prefixes=CHECK-FP16 %s
|
|
; RUN: %if ptxas-sm_90 && ptxas-isa-7.8 %{ llc < %s -mcpu=sm_90 -mattr=+ptx78 | %ptxas-verify -arch=sm_90 %}
|
|
target triple = "nvptx64-nvidia-cuda"
|
|
|
|
declare half @llvm.nvvm.ex2.approx.f16(half)
|
|
declare <2 x half> @llvm.nvvm.ex2.approx.v2f16(<2 x half>)
|
|
declare bfloat @llvm.nvvm.ex2.approx.ftz.bf16(bfloat)
|
|
declare <2 x bfloat> @llvm.nvvm.ex2.approx.ftz.v2bf16(<2 x bfloat>)
|
|
|
|
define half @ex2_half(half %0) {
|
|
; CHECK-FP16-LABEL: ex2_half(
|
|
; CHECK-FP16: {
|
|
; CHECK-FP16-NEXT: .reg .b16 %rs<3>;
|
|
; CHECK-FP16-EMPTY:
|
|
; CHECK-FP16-NEXT: // %bb.0:
|
|
; CHECK-FP16-NEXT: ld.param.b16 %rs1, [ex2_half_param_0];
|
|
; CHECK-FP16-NEXT: ex2.approx.f16 %rs2, %rs1;
|
|
; CHECK-FP16-NEXT: st.param.b16 [func_retval0], %rs2;
|
|
; CHECK-FP16-NEXT: ret;
|
|
%res = call half @llvm.nvvm.ex2.approx.f16(half %0)
|
|
ret half %res
|
|
}
|
|
|
|
define <2 x half> @ex2_2xhalf(<2 x half> %0) {
|
|
; CHECK-FP16-LABEL: ex2_2xhalf(
|
|
; CHECK-FP16: {
|
|
; CHECK-FP16-NEXT: .reg .b32 %r<3>;
|
|
; CHECK-FP16-EMPTY:
|
|
; CHECK-FP16-NEXT: // %bb.0:
|
|
; CHECK-FP16-NEXT: ld.param.b32 %r1, [ex2_2xhalf_param_0];
|
|
; CHECK-FP16-NEXT: ex2.approx.f16x2 %r2, %r1;
|
|
; CHECK-FP16-NEXT: st.param.b32 [func_retval0], %r2;
|
|
; CHECK-FP16-NEXT: ret;
|
|
%res = call <2 x half> @llvm.nvvm.ex2.approx.v2f16(<2 x half> %0)
|
|
ret <2 x half> %res
|
|
}
|
|
|
|
define bfloat @ex2_bfloat(bfloat %0) {
|
|
; CHECK-FP16-LABEL: ex2_bfloat(
|
|
; CHECK-FP16: {
|
|
; CHECK-FP16-NEXT: .reg .b16 %rs<3>;
|
|
; CHECK-FP16-EMPTY:
|
|
; CHECK-FP16-NEXT: // %bb.0:
|
|
; CHECK-FP16-NEXT: ld.param.b16 %rs1, [ex2_bfloat_param_0];
|
|
; CHECK-FP16-NEXT: ex2.approx.ftz.bf16 %rs2, %rs1;
|
|
; CHECK-FP16-NEXT: st.param.b16 [func_retval0], %rs2;
|
|
; CHECK-FP16-NEXT: ret;
|
|
%res = call bfloat @llvm.nvvm.ex2.approx.ftz.bf16(bfloat %0)
|
|
ret bfloat %res
|
|
}
|
|
|
|
define <2 x bfloat> @ex2_2xbfloat(<2 x bfloat> %0) {
|
|
; CHECK-FP16-LABEL: ex2_2xbfloat(
|
|
; CHECK-FP16: {
|
|
; CHECK-FP16-NEXT: .reg .b32 %r<3>;
|
|
; CHECK-FP16-EMPTY:
|
|
; CHECK-FP16-NEXT: // %bb.0:
|
|
; CHECK-FP16-NEXT: ld.param.b32 %r1, [ex2_2xbfloat_param_0];
|
|
; CHECK-FP16-NEXT: ex2.approx.ftz.bf16x2 %r2, %r1;
|
|
; CHECK-FP16-NEXT: st.param.b32 [func_retval0], %r2;
|
|
; CHECK-FP16-NEXT: ret;
|
|
%res = call <2 x bfloat> @llvm.nvvm.ex2.approx.ftz.v2bf16(<2 x bfloat> %0)
|
|
ret <2 x bfloat> %res
|
|
}
|