31 llvm/test/CodeGen/AMDGPU/fp-atomics-gfx940.ll
View file
@@ -31,12 +31,11 @@ define amdgpu_kernel void @flat_atomic_fadd_f32_noret_pat(float* %ptr) {
; GFX940-NEXT: v_mov_b32_e32 v2, 4.0
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: v_mov_b64_e32 v[0:1], s[0:1]
; GFX940-NEXT: buffer_wbl2
; GFX940-NEXT: buffer_wbl2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: flat_atomic_add_f32 v[0:1], v2
; GFX940-NEXT: flat_atomic_add_f32 v[0:1], v2 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_invl2
; GFX940-NEXT: buffer_wbinvl1_vol
; GFX940-NEXT: buffer_inv sc0 sc1
; GFX940-NEXT: s_endpgm
%ret = atomicrmw fadd float* %ptr, float 4.0 seq_cst
ret void
@@ -49,12 +48,11 @@ define amdgpu_kernel void @flat_atomic_fadd_f32_noret_pat_ieee(float* %ptr) #0 {
; GFX940-NEXT: v_mov_b32_e32 v2, 4.0
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: v_mov_b64_e32 v[0:1], s[0:1]
; GFX940-NEXT: buffer_wbl2
; GFX940-NEXT: buffer_wbl2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: flat_atomic_add_f32 v[0:1], v2
; GFX940-NEXT: flat_atomic_add_f32 v[0:1], v2 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_invl2
; GFX940-NEXT: buffer_wbinvl1_vol
; GFX940-NEXT: buffer_inv sc0 sc1
; GFX940-NEXT: s_endpgm
%ret = atomicrmw fadd float* %ptr, float 4.0 seq_cst
ret void
@@ -76,12 +74,11 @@ define float @flat_atomic_fadd_f32_rtn_pat(float* %ptr, float %data) {
; GFX940: ; %bb.0:
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v2, 4.0
; GFX940-NEXT: buffer_wbl2
; GFX940-NEXT: buffer_wbl2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: flat_atomic_add_f32 v0, v[0:1], v2 sc0
; GFX940-NEXT: flat_atomic_add_f32 v0, v[0:1], v2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_invl2
; GFX940-NEXT: buffer_wbinvl1_vol
; GFX940-NEXT: buffer_inv sc0 sc1
; GFX940-NEXT: s_setpc_b64 s[30:31]
%ret = atomicrmw fadd float* %ptr, float 4.0 seq_cst
ret float %ret
@@ -195,12 +192,11 @@ define amdgpu_kernel void @local_atomic_fadd_v2bf16_noret(<2 x i16> addrspace(3)
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, s2
; GFX940-NEXT: v_mov_b32_e32 v1, s3
; GFX940-NEXT: buffer_wbl2
; GFX940-NEXT: buffer_wbl2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: ds_pk_add_bf16 v0, v1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_invl2
; GFX940-NEXT: buffer_wbinvl1_vol
; GFX940-NEXT: buffer_inv sc0 sc1
; GFX940-NEXT: s_endpgm
%ret = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(<2 x i16> addrspace(3)* %ptr, <2 x i16> %data)
ret void
@@ -210,12 +206,11 @@ define <2 x i16> @local_atomic_fadd_v2bf16_rtn(<2 x i16> addrspace(3)* %ptr, <2
; GFX940-LABEL: local_atomic_fadd_v2bf16_rtn:
; GFX940: ; %bb.0:
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_wbl2
; GFX940-NEXT: buffer_wbl2 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: ds_pk_add_rtn_bf16 v0, v0, v1
; GFX940-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NEXT: buffer_invl2
; GFX940-NEXT: buffer_wbinvl1_vol
; GFX940-NEXT: buffer_inv sc0 sc1
; GFX940-NEXT: s_setpc_b64 s[30:31]
%ret = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(<2 x i16> addrspace(3)* %ptr, <2 x i16> %data)
ret <2 x i16> %ret
428 llvm/test/CodeGen/AMDGPU/memory-legalizer-fence.ll
View file

Large diffs are not rendered by default.

2,488 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-agent.ll
View file

Large diffs are not rendered by default.

114 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-nontemporal.ll
View file
@@ -5,6 +5,8 @@
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx700 -amdgcn-skip-cache-invalidations -verify-machineinstrs < %s | FileCheck --check-prefixes=SKIP-CACHE-INV %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-TGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-TGSPLIT %s

define amdgpu_kernel void @flat_nontemporal_load_0(
; GFX7-LABEL: flat_nontemporal_load_0:
@@ -84,6 +86,32 @@ define amdgpu_kernel void @flat_nontemporal_load_0(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: flat_nontemporal_load_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-NOTTGSPLIT-NEXT: flat_load_dword v2, v[0:1] nt
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: flat_nontemporal_load_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-TGSPLIT-NEXT: flat_load_dword v2, v[0:1] nt
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX940-TGSPLIT-NEXT: s_endpgm
i32* %in, i32* %out) {
entry:
%val = load i32, i32* %in, align 4, !nontemporal !0
@@ -179,6 +207,36 @@ define amdgpu_kernel void @flat_nontemporal_load_1(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: flat_nontemporal_load_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-NOTTGSPLIT-NEXT: v_add_co_u32_e32 v0, vcc, s0, v0
; GFX940-NOTTGSPLIT-NEXT: v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX940-NOTTGSPLIT-NEXT: flat_load_dword v2, v[0:1] nt
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: flat_nontemporal_load_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-TGSPLIT-NEXT: v_add_co_u32_e32 v0, vcc, s0, v0
; GFX940-TGSPLIT-NEXT: v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX940-TGSPLIT-NEXT: flat_load_dword v2, v[0:1] nt
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: flat_store_dword v[0:1], v2
; GFX940-TGSPLIT-NEXT: s_endpgm
i32* %in, i32* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -266,6 +324,32 @@ define amdgpu_kernel void @flat_nontemporal_store_0(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: flat_store_dword v[0:1], v2 glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: flat_nontemporal_store_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-NOTTGSPLIT-NEXT: flat_load_dword v2, v[0:1]
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: flat_store_dword v[0:1], v2 nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: flat_nontemporal_store_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s1
; GFX940-TGSPLIT-NEXT: flat_load_dword v2, v[0:1]
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s2
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: flat_store_dword v[0:1], v2 nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32* %in, i32* %out) {
entry:
%val = load i32, i32* %in, align 4
@@ -361,6 +445,36 @@ define amdgpu_kernel void @flat_nontemporal_store_1(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: flat_store_dword v[0:1], v2 glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: flat_nontemporal_store_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v2, s0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v3, s1
; GFX940-NOTTGSPLIT-NEXT: flat_load_dword v2, v[2:3]
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-NOTTGSPLIT-NEXT: v_add_co_u32_e32 v0, vcc, s2, v0
; GFX940-NOTTGSPLIT-NEXT: v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: flat_store_dword v[0:1], v2 nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: flat_nontemporal_store_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v2, s0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v3, s1
; GFX940-TGSPLIT-NEXT: flat_load_dword v2, v[2:3]
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s3
; GFX940-TGSPLIT-NEXT: v_add_co_u32_e32 v0, vcc, s2, v0
; GFX940-TGSPLIT-NEXT: v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: flat_store_dword v[0:1], v2 nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32* %in, i32* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
2,034 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-singlethread.ll
View file

Large diffs are not rendered by default.

2,488 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-system.ll
View file

Large diffs are not rendered by default.

2,010 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-wavefront.ll
View file

Large diffs are not rendered by default.

2,142 llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll
View file

Large diffs are not rendered by default.

2,390 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-agent.ll
View file

Large diffs are not rendered by default.

90 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-nontemporal.ll
View file
@@ -6,6 +6,8 @@
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx700 -amdgcn-skip-cache-invalidations -verify-machineinstrs < %s | FileCheck --check-prefixes=SKIP-CACHE-INV %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-TGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-TGSPLIT %s

define amdgpu_kernel void @global_nontemporal_load_0(
; GFX6-LABEL: global_nontemporal_load_0:
@@ -91,6 +93,28 @@ define amdgpu_kernel void @global_nontemporal_load_0(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: global_nontemporal_load_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v0, v1, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: global_nontemporal_load_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
entry:
%val = load i32, i32 addrspace(1)* %in, align 4, !nontemporal !0
@@ -191,6 +215,28 @@ define amdgpu_kernel void @global_nontemporal_load_1(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: global_nontemporal_load_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_load_dword v0, v0, s[0:1] nt
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: global_nontemporal_load_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: global_load_dword v0, v0, s[0:1] nt
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -284,6 +330,28 @@ define amdgpu_kernel void @global_nontemporal_store_0(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: global_nontemporal_store_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: global_nontemporal_store_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
entry:
%val = load i32, i32 addrspace(1)* %in, align 4
@@ -379,6 +447,28 @@ define amdgpu_kernel void @global_nontemporal_store_1(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: global_nontemporal_store_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: global_nontemporal_store_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: global_store_dword v0, v1, s[2:3] nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(1)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
1,986 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-singlethread.ll
View file

Large diffs are not rendered by default.

2,278 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-system.ll
View file

Large diffs are not rendered by default.

1,986 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-wavefront.ll
View file

Large diffs are not rendered by default.

2,173 llvm/test/CodeGen/AMDGPU/memory-legalizer-global-workgroup.ll
View file

Large diffs are not rendered by default.

2,064 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-agent.ll
View file

Large diffs are not rendered by default.

98 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-nontemporal.ll
View file
@@ -6,6 +6,8 @@
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx700 -amdgcn-skip-cache-invalidations -verify-machineinstrs < %s | FileCheck --check-prefixes=SKIP-CACHE-INV %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-TGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-TGSPLIT %s

define amdgpu_kernel void @local_nontemporal_load_0(
; GFX6-LABEL: local_nontemporal_load_0:
@@ -99,6 +101,30 @@ define amdgpu_kernel void @local_nontemporal_load_0(
; GFX90A-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX90A-TGSPLIT-NEXT: global_store_dword v1, v0, s[0:1]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: local_nontemporal_load_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s4
; GFX940-NOTTGSPLIT-NEXT: ds_read_b32 v0, v0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: local_nontemporal_load_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s4
; GFX940-TGSPLIT-NEXT: ds_read_b32 v0, v0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(3)* %in, i32 addrspace(1)* %out) {
entry:
%val = load i32, i32 addrspace(3)* %in, align 4, !nontemporal !0
@@ -201,6 +227,30 @@ define amdgpu_kernel void @local_nontemporal_load_1(
; GFX90A-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX90A-TGSPLIT-NEXT: global_store_dword v1, v0, s[0:1]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: local_nontemporal_load_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-NOTTGSPLIT-NEXT: ds_read_b32 v0, v0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: local_nontemporal_load_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-TGSPLIT-NEXT: ds_read_b32 v0, v0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(3)* %in, i32 addrspace(1)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -299,6 +349,30 @@ define amdgpu_kernel void @local_nontemporal_store_0(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: local_nontemporal_store_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s4
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: local_nontemporal_store_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s4
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(3)* %out) {
entry:
%val = load i32, i32 addrspace(1)* %in, align 4
@@ -398,6 +472,30 @@ define amdgpu_kernel void @local_nontemporal_store_1(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: local_nontemporal_store_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: local_nontemporal_store_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: ds_write_b32 v0, v1
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(3)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
1,954 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-singlethread.ll
View file

Large diffs are not rendered by default.

2,064 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-system.ll
View file

Large diffs are not rendered by default.

1,954 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-wavefront.ll
View file

Large diffs are not rendered by default.

2,064 llvm/test/CodeGen/AMDGPU/memory-legalizer-local-workgroup.ll
View file

Large diffs are not rendered by default.

94 llvm/test/CodeGen/AMDGPU/memory-legalizer-private-nontemporal.ll
View file
@@ -6,6 +6,8 @@
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx700 -amdgcn-skip-cache-invalidations -verify-machineinstrs < %s | FileCheck --check-prefixes=SKIP-CACHE-INV %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX90A-TGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-NOTTGSPLIT %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx940 -mattr=+tgsplit -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX940-TGSPLIT %s

define amdgpu_kernel void @private_nontemporal_load_0(
; GFX6-LABEL: private_nontemporal_load_0:
@@ -125,6 +127,28 @@ define amdgpu_kernel void @private_nontemporal_load_0(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: global_store_dword v1, v0, s[0:1]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: private_nontemporal_load_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: scratch_load_dword v0, off, s4 nt
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: private_nontemporal_load_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: scratch_load_dword v0, off, s4 nt
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(5)* %in, i32 addrspace(1)* %out) {
entry:
%val = load i32, i32 addrspace(5)* %in, align 4, !nontemporal !0
@@ -253,6 +277,30 @@ define amdgpu_kernel void @private_nontemporal_load_1(
; GFX90A-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX90A-TGSPLIT-NEXT: global_store_dword v1, v0, s[0:1]
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: private_nontemporal_load_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-NOTTGSPLIT-NEXT: scratch_load_dword v0, v0, off nt
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: private_nontemporal_load_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, 0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-TGSPLIT-NEXT: scratch_load_dword v0, v0, off nt
; GFX940-TGSPLIT-NEXT: s_waitcnt vmcnt(0)
; GFX940-TGSPLIT-NEXT: global_store_dword v1, v0, s[2:3]
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(5)* %in, i32 addrspace(1)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()
@@ -378,6 +426,28 @@ define amdgpu_kernel void @private_nontemporal_store_0(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX90A-TGSPLIT-NEXT: buffer_store_dword v0, v1, s[8:11], 0 offen glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: private_nontemporal_store_0:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NOTTGSPLIT-NEXT: scratch_store_dword off, v0, s4 nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: private_nontemporal_store_0:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v0, s0
; GFX940-TGSPLIT-NEXT: scratch_store_dword off, v0, s4 nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(5)* %out) {
entry:
%val = load i32, i32 addrspace(1)* %in, align 4
@@ -504,6 +574,30 @@ define amdgpu_kernel void @private_nontemporal_store_1(
; GFX90A-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX90A-TGSPLIT-NEXT: buffer_store_dword v1, v0, s[8:11], 0 offen glc slc
; GFX90A-TGSPLIT-NEXT: s_endpgm
;
; GFX940-NOTTGSPLIT-LABEL: private_nontemporal_store_1:
; GFX940-NOTTGSPLIT: ; %bb.0: ; %entry
; GFX940-NOTTGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-NOTTGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-NOTTGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NOTTGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-NOTTGSPLIT-NEXT: scratch_store_dword v0, v1, off nt
; GFX940-NOTTGSPLIT-NEXT: s_endpgm
;
; GFX940-TGSPLIT-LABEL: private_nontemporal_store_1:
; GFX940-TGSPLIT: ; %bb.0: ; %entry
; GFX940-TGSPLIT-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0x0
; GFX940-TGSPLIT-NEXT: s_load_dword s4, s[0:1], 0x8
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_lshl_add_u32 v0, v0, 2, s4
; GFX940-TGSPLIT-NEXT: s_load_dword s0, s[2:3], 0x0
; GFX940-TGSPLIT-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-TGSPLIT-NEXT: v_mov_b32_e32 v1, s0
; GFX940-TGSPLIT-NEXT: scratch_store_dword v0, v1, off nt
; GFX940-TGSPLIT-NEXT: s_endpgm
i32 addrspace(1)* %in, i32 addrspace(5)* %out) {
entry:
%tid = call i32 @llvm.amdgcn.workitem.id.x()