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
584 lines
22 KiB
LLVM
584 lines
22 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 < %s | FileCheck --check-prefix=GFX9 %s
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=hawaii < %s | FileCheck --check-prefix=GFX7 %s
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 < %s | FileCheck --check-prefix=GFX10 %s
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx1100 < %s | FileCheck --check-prefix=GFX11 %s
|
|
|
|
; FIXME:
|
|
; XUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=tahiti < %s | FileCheck --check-prefix=GFX6 %s
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX9-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX9-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX9-NEXT: v_mov_b32_e32 v4, s6
|
|
; GFX9-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX7-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX7-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX7-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX7-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s6
|
|
; GFX10-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, s1
|
|
; GFX11-NEXT: v_dual_mov_b32 v2, s2 :: v_dual_mov_b32 v3, s3
|
|
; GFX11-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX11-NEXT: ds_store_b128 v4, v[0:3]
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32_align1(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32_align1:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: s_and_b32 s5, 0xffff, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX9-NEXT: s_lshr_b32 s5, s5, 8
|
|
; GFX9-NEXT: s_lshr_b32 s4, s0, 16
|
|
; GFX9-NEXT: ds_write_b8 v1, v0
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s5
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:1
|
|
; GFX9-NEXT: s_lshr_b32 s0, s0, 24
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s4
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:2
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: s_and_b32 s4, 0xffff, s1
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:3
|
|
; GFX9-NEXT: s_lshr_b32 s4, s4, 8
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:4
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s4
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:5
|
|
; GFX9-NEXT: s_lshr_b32 s1, s1, 24
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:6
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: s_and_b32 s1, 0xffff, s2
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:7
|
|
; GFX9-NEXT: s_lshr_b32 s1, s1, 8
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s2
|
|
; GFX9-NEXT: s_lshr_b32 s0, s2, 16
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:8
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:9
|
|
; GFX9-NEXT: s_lshr_b32 s1, s2, 24
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:10
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: s_and_b32 s1, 0xffff, s3
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:11
|
|
; GFX9-NEXT: s_lshr_b32 s1, s1, 8
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s3
|
|
; GFX9-NEXT: s_lshr_b32 s0, s3, 16
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:12
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:13
|
|
; GFX9-NEXT: s_lshr_b32 s1, s3, 24
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:14
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: ds_write_b8 v1, v0 offset:15
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32_align1:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: s_bfe_u32 s6, s0, 0x80008
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s4
|
|
; GFX7-NEXT: s_lshr_b32 s5, s0, 16
|
|
; GFX7-NEXT: ds_write_b8 v1, v0
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s6
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:1
|
|
; GFX7-NEXT: s_lshr_b32 s0, s0, 24
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s5
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:2
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:3
|
|
; GFX7-NEXT: s_bfe_u32 s4, s1, 0x80008
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:4
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s4
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:5
|
|
; GFX7-NEXT: s_lshr_b32 s1, s1, 24
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:6
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:7
|
|
; GFX7-NEXT: s_bfe_u32 s1, s2, 0x80008
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s2
|
|
; GFX7-NEXT: s_lshr_b32 s0, s2, 16
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:8
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:9
|
|
; GFX7-NEXT: s_lshr_b32 s1, s2, 24
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:10
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:11
|
|
; GFX7-NEXT: s_bfe_u32 s1, s3, 0x80008
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s3
|
|
; GFX7-NEXT: s_lshr_b32 s0, s3, 16
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:12
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:13
|
|
; GFX7-NEXT: s_lshr_b32 s1, s3, 24
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:14
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b8 v1, v0 offset:15
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32_align1:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: s_lshr_b32 s4, s0, 16
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX10-NEXT: s_and_b32 s5, 0xffff, s0
|
|
; GFX10-NEXT: s_lshr_b32 s0, s0, 24
|
|
; GFX10-NEXT: s_and_b32 s7, 0xffff, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX10-NEXT: s_lshr_b32 s6, s1, 16
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX10-NEXT: s_lshr_b32 s8, s2, 16
|
|
; GFX10-NEXT: s_lshr_b32 s5, s5, 8
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, s0
|
|
; GFX10-NEXT: s_lshr_b32 s0, s7, 8
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, s6
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, s5
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, s0
|
|
; GFX10-NEXT: ds_write_b8 v1, v0
|
|
; GFX10-NEXT: ds_write_b8 v1, v2 offset:4
|
|
; GFX10-NEXT: ds_write_b8 v1, v4 offset:2
|
|
; GFX10-NEXT: ds_write_b8 v1, v5 offset:3
|
|
; GFX10-NEXT: ds_write_b8 v1, v6 offset:6
|
|
; GFX10-NEXT: ds_write_b8 v1, v8 offset:1
|
|
; GFX10-NEXT: ds_write_b8 v1, v9 offset:5
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s8
|
|
; GFX10-NEXT: s_lshr_b32 s1, s1, 24
|
|
; GFX10-NEXT: s_and_b32 s9, 0xffff, s2
|
|
; GFX10-NEXT: s_lshr_b32 s0, s2, 24
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, s1
|
|
; GFX10-NEXT: s_lshr_b32 s1, s9, 8
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s2
|
|
; GFX10-NEXT: ds_write_b8 v1, v0 offset:10
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: s_and_b32 s0, 0xffff, s3
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, s1
|
|
; GFX10-NEXT: s_lshr_b32 s0, s0, 8
|
|
; GFX10-NEXT: s_lshr_b32 s1, s3, 16
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s3
|
|
; GFX10-NEXT: ds_write_b8 v1, v7 offset:7
|
|
; GFX10-NEXT: ds_write_b8 v1, v3 offset:8
|
|
; GFX10-NEXT: ds_write_b8 v1, v10 offset:9
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s0
|
|
; GFX10-NEXT: s_lshr_b32 s0, s3, 24
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, s0
|
|
; GFX10-NEXT: ds_write_b8 v1, v0 offset:11
|
|
; GFX10-NEXT: ds_write_b8 v1, v2 offset:12
|
|
; GFX10-NEXT: ds_write_b8 v1, v3 offset:13
|
|
; GFX10-NEXT: ds_write_b8 v1, v4 offset:14
|
|
; GFX10-NEXT: ds_write_b8 v1, v5 offset:15
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32_align1:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: s_and_b32 s6, 0xffff, s0
|
|
; GFX11-NEXT: s_lshr_b32 s5, s0, 16
|
|
; GFX11-NEXT: v_mov_b16_e32 v0.l, s0
|
|
; GFX11-NEXT: s_lshr_b32 s0, s0, 24
|
|
; GFX11-NEXT: s_and_b32 s7, 0xffff, s1
|
|
; GFX11-NEXT: s_lshr_b32 s6, s6, 8
|
|
; GFX11-NEXT: v_mov_b32_e32 v5, s4
|
|
; GFX11-NEXT: v_mov_b16_e32 v2.l, s0
|
|
; GFX11-NEXT: s_lshr_b32 s0, s7, 8
|
|
; GFX11-NEXT: v_mov_b16_e32 v3.l, s6
|
|
; GFX11-NEXT: v_mov_b16_e32 v1.h, s5
|
|
; GFX11-NEXT: s_lshr_b32 s4, s1, 16
|
|
; GFX11-NEXT: v_mov_b16_e32 v0.h, s1
|
|
; GFX11-NEXT: s_lshr_b32 s1, s1, 24
|
|
; GFX11-NEXT: s_and_b32 s9, 0xffff, s2
|
|
; GFX11-NEXT: v_mov_b16_e32 v4.l, s0
|
|
; GFX11-NEXT: s_lshr_b32 s0, s2, 24
|
|
; GFX11-NEXT: v_mov_b16_e32 v2.h, s4
|
|
; GFX11-NEXT: v_mov_b16_e32 v3.h, s1
|
|
; GFX11-NEXT: s_lshr_b32 s1, s9, 8
|
|
; GFX11-NEXT: ds_store_b8 v5, v0
|
|
; GFX11-NEXT: ds_store_b8 v5, v3 offset:1
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v1 offset:2
|
|
; GFX11-NEXT: ds_store_b8 v5, v2 offset:3
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v0 offset:4
|
|
; GFX11-NEXT: ds_store_b8 v5, v4 offset:5
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v2 offset:6
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v3 offset:7
|
|
; GFX11-NEXT: v_mov_b16_e32 v1.h, s0
|
|
; GFX11-NEXT: s_and_b32 s0, 0xffff, s3
|
|
; GFX11-NEXT: s_lshr_b32 s8, s2, 16
|
|
; GFX11-NEXT: v_mov_b16_e32 v1.l, s2
|
|
; GFX11-NEXT: v_mov_b16_e32 v0.l, s1
|
|
; GFX11-NEXT: s_lshr_b32 s0, s0, 8
|
|
; GFX11-NEXT: v_mov_b16_e32 v0.h, s8
|
|
; GFX11-NEXT: s_lshr_b32 s1, s3, 16
|
|
; GFX11-NEXT: v_mov_b16_e32 v2.h, s0
|
|
; GFX11-NEXT: s_lshr_b32 s0, s3, 24
|
|
; GFX11-NEXT: v_mov_b16_e32 v2.l, s3
|
|
; GFX11-NEXT: v_mov_b16_e32 v3.l, s1
|
|
; GFX11-NEXT: v_mov_b16_e32 v3.h, s0
|
|
; GFX11-NEXT: ds_store_b8 v5, v1 offset:8
|
|
; GFX11-NEXT: ds_store_b8 v5, v0 offset:9
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v0 offset:10
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v1 offset:11
|
|
; GFX11-NEXT: ds_store_b8 v5, v2 offset:12
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v2 offset:13
|
|
; GFX11-NEXT: ds_store_b8 v5, v3 offset:14
|
|
; GFX11-NEXT: ds_store_b8_d16_hi v5, v3 offset:15
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out, align 1
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32_align2(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32_align2:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: s_lshr_b32 s4, s0, 16
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX9-NEXT: ds_write_b16 v1, v0
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s4
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:2
|
|
; GFX9-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:4
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:6
|
|
; GFX9-NEXT: s_lshr_b32 s0, s2, 16
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s2
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:8
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:10
|
|
; GFX9-NEXT: s_lshr_b32 s0, s3, 16
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s3
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:12
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: ds_write_b16 v1, v0 offset:14
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32_align2:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: s_lshr_b32 s5, s0, 16
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s4
|
|
; GFX7-NEXT: ds_write_b16 v1, v0
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s5
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:2
|
|
; GFX7-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s1
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:4
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:6
|
|
; GFX7-NEXT: s_lshr_b32 s0, s2, 16
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s2
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:8
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:10
|
|
; GFX7-NEXT: s_lshr_b32 s0, s3, 16
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s3
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:12
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: ds_write_b16 v1, v0 offset:14
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32_align2:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX10-NEXT: s_lshr_b32 s4, s0, 16
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s2
|
|
; GFX10-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX10-NEXT: s_lshr_b32 s1, s2, 16
|
|
; GFX10-NEXT: s_lshr_b32 s2, s3, 16
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s3
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, s4
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, s2
|
|
; GFX10-NEXT: ds_write_b16 v1, v0
|
|
; GFX10-NEXT: ds_write_b16 v1, v2 offset:4
|
|
; GFX10-NEXT: ds_write_b16 v1, v3 offset:8
|
|
; GFX10-NEXT: ds_write_b16 v1, v4 offset:12
|
|
; GFX10-NEXT: ds_write_b16 v1, v5 offset:2
|
|
; GFX10-NEXT: ds_write_b16 v1, v6 offset:6
|
|
; GFX10-NEXT: ds_write_b16 v1, v7 offset:10
|
|
; GFX10-NEXT: ds_write_b16 v1, v8 offset:14
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32_align2:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: s_lshr_b32 s5, s0, 16
|
|
; GFX11-NEXT: v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, s4
|
|
; GFX11-NEXT: s_lshr_b32 s0, s1, 16
|
|
; GFX11-NEXT: v_dual_mov_b32 v4, s3 :: v_dual_mov_b32 v5, s5
|
|
; GFX11-NEXT: v_dual_mov_b32 v2, s1 :: v_dual_mov_b32 v3, s2
|
|
; GFX11-NEXT: s_lshr_b32 s1, s2, 16
|
|
; GFX11-NEXT: s_lshr_b32 s2, s3, 16
|
|
; GFX11-NEXT: v_dual_mov_b32 v6, s0 :: v_dual_mov_b32 v7, s1
|
|
; GFX11-NEXT: v_mov_b32_e32 v8, s2
|
|
; GFX11-NEXT: ds_store_b16 v1, v0
|
|
; GFX11-NEXT: ds_store_b16 v1, v5 offset:2
|
|
; GFX11-NEXT: ds_store_b16 v1, v2 offset:4
|
|
; GFX11-NEXT: ds_store_b16 v1, v6 offset:6
|
|
; GFX11-NEXT: ds_store_b16 v1, v3 offset:8
|
|
; GFX11-NEXT: ds_store_b16 v1, v7 offset:10
|
|
; GFX11-NEXT: ds_store_b16 v1, v4 offset:12
|
|
; GFX11-NEXT: ds_store_b16 v1, v8 offset:14
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out, align 2
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32_align4(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32_align4:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX9-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX9-NEXT: v_mov_b32_e32 v3, s2
|
|
; GFX9-NEXT: ds_write2_b32 v1, v0, v2 offset1:1
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s3
|
|
; GFX9-NEXT: ds_write2_b32 v1, v3, v0 offset0:2 offset1:3
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32_align4:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s4
|
|
; GFX7-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX7-NEXT: ds_write2_b32 v1, v0, v2 offset1:1
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s2
|
|
; GFX7-NEXT: v_mov_b32_e32 v2, s3
|
|
; GFX7-NEXT: ds_write2_b32 v1, v0, v2 offset0:2 offset1:3
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32_align4:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s2
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s3
|
|
; GFX10-NEXT: ds_write2_b32 v1, v0, v2 offset1:1
|
|
; GFX10-NEXT: ds_write2_b32 v1, v3, v4 offset0:2 offset1:3
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32_align4:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, s4
|
|
; GFX11-NEXT: v_dual_mov_b32 v2, s1 :: v_dual_mov_b32 v3, s2
|
|
; GFX11-NEXT: v_mov_b32_e32 v4, s3
|
|
; GFX11-NEXT: ds_store_2addr_b32 v1, v0, v2 offset1:1
|
|
; GFX11-NEXT: ds_store_2addr_b32 v1, v3, v4 offset0:2 offset1:3
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out, align 4
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32_align8(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32_align8:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX9-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX9-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX9-NEXT: v_mov_b32_e32 v4, s6
|
|
; GFX9-NEXT: ds_write2_b64 v4, v[0:1], v[2:3] offset1:1
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32_align8:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX7-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX7-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX7-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX7-NEXT: ds_write2_b64 v4, v[0:1], v[2:3] offset1:1
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32_align8:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s6
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s2
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s3
|
|
; GFX10-NEXT: ds_write2_b32 v1, v0, v2 offset1:1
|
|
; GFX10-NEXT: ds_write2_b32 v1, v3, v4 offset0:2 offset1:3
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32_align8:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, s1
|
|
; GFX11-NEXT: v_dual_mov_b32 v2, s2 :: v_dual_mov_b32 v3, s3
|
|
; GFX11-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX11-NEXT: ds_store_2addr_b64 v4, v[0:1], v[2:3] offset1:1
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out, align 8
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @store_lds_v4i32_align16(ptr addrspace(3) %out, <4 x i32> %x) {
|
|
; GFX9-LABEL: store_lds_v4i32_align16:
|
|
; GFX9: ; %bb.0:
|
|
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX9-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX9-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX9-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX9-NEXT: v_mov_b32_e32 v4, s6
|
|
; GFX9-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX9-NEXT: s_endpgm
|
|
;
|
|
; GFX7-LABEL: store_lds_v4i32_align16:
|
|
; GFX7: ; %bb.0:
|
|
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x4
|
|
; GFX7-NEXT: s_load_dword s4, s[4:5], 0x0
|
|
; GFX7-NEXT: s_mov_b32 m0, -1
|
|
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX7-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX7-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX7-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX7-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX7-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX7-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX7-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_lds_v4i32_align16:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: s_clause 0x1
|
|
; GFX10-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x10
|
|
; GFX10-NEXT: s_load_dword s6, s[4:5], 0x0
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, s1
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, s2
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, s3
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, s6
|
|
; GFX10-NEXT: ds_write_b128 v4, v[0:3]
|
|
; GFX10-NEXT: s_endpgm
|
|
;
|
|
; GFX11-LABEL: store_lds_v4i32_align16:
|
|
; GFX11: ; %bb.0:
|
|
; GFX11-NEXT: s_clause 0x1
|
|
; GFX11-NEXT: s_load_b128 s[0:3], s[4:5], 0x10
|
|
; GFX11-NEXT: s_load_b32 s4, s[4:5], 0x0
|
|
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX11-NEXT: v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, s1
|
|
; GFX11-NEXT: v_dual_mov_b32 v2, s2 :: v_dual_mov_b32 v3, s3
|
|
; GFX11-NEXT: v_mov_b32_e32 v4, s4
|
|
; GFX11-NEXT: ds_store_b128 v4, v[0:3]
|
|
; GFX11-NEXT: s_endpgm
|
|
store <4 x i32> %x, ptr addrspace(3) %out, align 16
|
|
ret void
|
|
}
|