Files
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

109 lines
4.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_20 | FileCheck %s
; RUN: %if ptxas %{ llc < %s -mtriple=nvptx64 -mcpu=sm_20 | %ptxas-verify %}
target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v16:16:16-v32:32:32-v64:64:64-v128:128:128-n16:32:64"
define <16 x float> @test_v16f32(<16 x float> %a) {
; CHECK-LABEL: test_v16f32(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<17>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v4.b32 {%r1, %r2, %r3, %r4}, [test_v16f32_param_0];
; CHECK-NEXT: ld.param.v4.b32 {%r5, %r6, %r7, %r8}, [test_v16f32_param_0+16];
; CHECK-NEXT: ld.param.v4.b32 {%r9, %r10, %r11, %r12}, [test_v16f32_param_0+32];
; CHECK-NEXT: ld.param.v4.b32 {%r13, %r14, %r15, %r16}, [test_v16f32_param_0+48];
; CHECK-NEXT: st.param.v4.b32 [func_retval0+48], {%r13, %r14, %r15, %r16};
; CHECK-NEXT: st.param.v4.b32 [func_retval0+32], {%r9, %r10, %r11, %r12};
; CHECK-NEXT: st.param.v4.b32 [func_retval0+16], {%r5, %r6, %r7, %r8};
; CHECK-NEXT: st.param.v4.b32 [func_retval0], {%r1, %r2, %r3, %r4};
; CHECK-NEXT: ret;
ret <16 x float> %a
}
define <8 x float> @test_v8f32(<8 x float> %a) {
; CHECK-LABEL: test_v8f32(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<9>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v4.b32 {%r1, %r2, %r3, %r4}, [test_v8f32_param_0];
; CHECK-NEXT: ld.param.v4.b32 {%r5, %r6, %r7, %r8}, [test_v8f32_param_0+16];
; CHECK-NEXT: st.param.v4.b32 [func_retval0+16], {%r5, %r6, %r7, %r8};
; CHECK-NEXT: st.param.v4.b32 [func_retval0], {%r1, %r2, %r3, %r4};
; CHECK-NEXT: ret;
ret <8 x float> %a
}
define <4 x float> @test_v4f32(<4 x float> %a) {
; CHECK-LABEL: test_v4f32(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<5>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v4.b32 {%r1, %r2, %r3, %r4}, [test_v4f32_param_0];
; CHECK-NEXT: st.param.v4.b32 [func_retval0], {%r1, %r2, %r3, %r4};
; CHECK-NEXT: ret;
ret <4 x float> %a
}
define <2 x float> @test_v2f32(<2 x float> %a) {
; CHECK-LABEL: test_v2f32(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<3>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v2.b32 {%r1, %r2}, [test_v2f32_param_0];
; CHECK-NEXT: st.param.v2.b32 [func_retval0], {%r1, %r2};
; CHECK-NEXT: ret;
ret <2 x float> %a
}
; Oddly shaped vectors should not load any extra elements.
define <3 x float> @test_v3f32(<3 x float> %a) {
; CHECK-LABEL: test_v3f32(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<4>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v2.b32 {%r1, %r2}, [test_v3f32_param_0];
; CHECK-NEXT: ld.param.b32 %r3, [test_v3f32_param_0+8];
; CHECK-NEXT: st.param.b32 [func_retval0+8], %r3;
; CHECK-NEXT: st.param.v2.b32 [func_retval0], {%r1, %r2};
; CHECK-NEXT: ret;
ret <3 x float> %a
}
define <8 x i64> @test_v8i64(<8 x i64> %a) {
; CHECK-LABEL: test_v8i64(
; CHECK: {
; CHECK-NEXT: .reg .b64 %rd<9>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v2.b64 {%rd1, %rd2}, [test_v8i64_param_0];
; CHECK-NEXT: ld.param.v2.b64 {%rd3, %rd4}, [test_v8i64_param_0+16];
; CHECK-NEXT: ld.param.v2.b64 {%rd5, %rd6}, [test_v8i64_param_0+32];
; CHECK-NEXT: ld.param.v2.b64 {%rd7, %rd8}, [test_v8i64_param_0+48];
; CHECK-NEXT: st.param.v2.b64 [func_retval0+48], {%rd7, %rd8};
; CHECK-NEXT: st.param.v2.b64 [func_retval0+32], {%rd5, %rd6};
; CHECK-NEXT: st.param.v2.b64 [func_retval0+16], {%rd3, %rd4};
; CHECK-NEXT: st.param.v2.b64 [func_retval0], {%rd1, %rd2};
; CHECK-NEXT: ret;
ret <8 x i64> %a
}
define <16 x i16> @test_v16i16(<16 x i16> %a) {
; CHECK-LABEL: test_v16i16(
; CHECK: {
; CHECK-NEXT: .reg .b32 %r<9>;
; CHECK-EMPTY:
; CHECK-NEXT: // %bb.0:
; CHECK-NEXT: ld.param.v4.b32 {%r1, %r2, %r3, %r4}, [test_v16i16_param_0];
; CHECK-NEXT: ld.param.v4.b32 {%r5, %r6, %r7, %r8}, [test_v16i16_param_0+16];
; CHECK-NEXT: st.param.v4.b32 [func_retval0+16], {%r5, %r6, %r7, %r8};
; CHECK-NEXT: st.param.v4.b32 [func_retval0], {%r1, %r2, %r3, %r4};
; CHECK-NEXT: ret;
ret <16 x i16> %a
}