diff options
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/debugger-insert-nops.ll | 24 | ||||
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll | 6 | ||||
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/merge-stores.ll | 24 | ||||
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/private-element-size.ll | 12 | ||||
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll | 17 |
5 files changed, 47 insertions, 36 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/debugger-insert-nops.ll b/llvm/test/CodeGen/AMDGPU/debugger-insert-nops.ll index 7be7d9486a4..6638f4e2582 100644 --- a/llvm/test/CodeGen/AMDGPU/debugger-insert-nops.ll +++ b/llvm/test/CodeGen/AMDGPU/debugger-insert-nops.ll @@ -1,21 +1,13 @@ -; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECK -; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECKNOP +; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s -; This test expects that we have one instance for each line in some order with "s_nop 0" instances after each. - -; Check that each line appears at least once -; CHECK-DAG: test01.cl:2:3 -; CHECK-DAG: test01.cl:3:3 -; CHECK-DAG: test01.cl:4:3 +; CHECK: test01.cl:2:{{[0-9]+}} +; CHECK-NEXT: s_nop 0 +; CHECK: test01.cl:3:{{[0-9]+}} +; CHECK-NEXT: s_nop 0 -; Check that each of each of the lines consists of the line output, followed by "s_nop 0" -; CHECKNOP: test01.cl:{{[234]}}:3 -; CHECKNOP-NEXT: s_nop 0 -; CHECKNOP: test01.cl:{{[234]}}:3 -; CHECKNOP-NEXT: s_nop 0 -; CHECKNOP: test01.cl:{{[234]}}:3 -; CHECKNOP-NEXT: s_nop 0 +; CHECK: test01.cl:4:{{[0-9]+}} +; CHECK-NEXT: s_nop 0 ; CHECK: test01.cl:5:{{[0-9]+}} ; CHECK-NEXT: s_nop 0 @@ -29,7 +21,7 @@ entry: call void @llvm.dbg.declare(metadata i32 addrspace(1)** %A.addr, metadata !17, metadata !18), !dbg !19 %0 = load i32 addrspace(1)*, i32 addrspace(1)** %A.addr, align 4, !dbg !20 %arrayidx = getelementptr inbounds i32, i32 addrspace(1)* %0, i32 0, !dbg !20 - store i32 1, i32 addrspace(1)* %arrayidx, align 4, !dbg !20 + store i32 1, i32 addrspace(1)* %arrayidx, align 4, !dbg !21 %1 = load i32 addrspace(1)*, i32 addrspace(1)** %A.addr, align 4, !dbg !22 %arrayidx1 = getelementptr inbounds i32, i32 addrspace(1)* %1, i32 1, !dbg !22 store i32 2, i32 addrspace(1)* %arrayidx1, align 4, !dbg !23 diff --git a/llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll b/llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll index f35bf1d546f..d7da673ee84 100644 --- a/llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll +++ b/llvm/test/CodeGen/AMDGPU/insert_vector_elt.ll @@ -255,9 +255,11 @@ define void @dynamic_insertelement_v2i8(<2 x i8> addrspace(1)* %out, <2 x i8> %a ; GCN: buffer_load_ubyte v{{[0-9]+}}, off ; GCN: buffer_load_ubyte v{{[0-9]+}}, off -; GCN-DAG: buffer_store_byte v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:4 -; GCN-DAG: buffer_store_byte v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:5 ; GCN-DAG: buffer_store_byte v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:6 +; GCN-DAG: buffer_store_byte v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:5 +; GCN-DAG: buffer_store_byte v{{[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:4 + +; GCN: buffer_store_byte v{{[0-9]+}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offen{{$}} ; GCN: buffer_load_ubyte ; GCN: buffer_load_ubyte diff --git a/llvm/test/CodeGen/AMDGPU/merge-stores.ll b/llvm/test/CodeGen/AMDGPU/merge-stores.ll index fd2d8ee1189..07104ebc8c9 100644 --- a/llvm/test/CodeGen/AMDGPU/merge-stores.ll +++ b/llvm/test/CodeGen/AMDGPU/merge-stores.ll @@ -1,5 +1,8 @@ -; RUN: llc -march=amdgcn -verify-machineinstrs -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-AA %s -; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-AA %s +; RUN: llc -march=amdgcn -verify-machineinstrs -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-NOAA %s +; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-NOAA %s + +; RUN: llc -march=amdgcn -verify-machineinstrs -combiner-alias-analysis -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-AA %s +; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs -combiner-alias-analysis -amdgpu-load-store-vectorizer=0 < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=GCN-AA %s ; This test is mostly to test DAG store merging, so disable the vectorizer. ; Run with devices with different unaligned load restrictions. @@ -147,7 +150,12 @@ define void @merge_global_store_4_constants_f32(float addrspace(1)* %out) #0 { } ; GCN-LABEL: {{^}}merge_global_store_4_constants_mixed_i32_f32: -; GCN-AA: buffer_store_dwordx4 v +; GCN-NOAA: buffer_store_dwordx4 v + +; GCN-AA: buffer_store_dwordx2 +; GCN-AA: buffer_store_dword v +; GCN-AA: buffer_store_dword v + ; GCN: s_endpgm define void @merge_global_store_4_constants_mixed_i32_f32(float addrspace(1)* %out) #0 { %out.gep.1 = getelementptr float, float addrspace(1)* %out, i32 1 @@ -466,9 +474,17 @@ define void @merge_global_store_4_adjacent_loads_i8_natural_align(i8 addrspace(1 ret void } +; This works once AA is enabled on the subtarget ; GCN-LABEL: {{^}}merge_global_store_4_vector_elts_loads_v4i32: ; GCN: buffer_load_dwordx4 [[LOAD:v\[[0-9]+:[0-9]+\]]] -; GCN: buffer_store_dwordx4 [[LOAD]] + +; GCN-NOAA: buffer_store_dword v +; GCN-NOAA: buffer_store_dword v +; GCN-NOAA: buffer_store_dword v +; GCN-NOAA: buffer_store_dword v + +; GCN-AA: buffer_store_dwordx4 [[LOAD]] + ; GCN: s_endpgm define void @merge_global_store_4_vector_elts_loads_v4i32(i32 addrspace(1)* %out, <4 x i32> addrspace(1)* %in) #0 { %out.gep.1 = getelementptr i32, i32 addrspace(1)* %out, i32 1 diff --git a/llvm/test/CodeGen/AMDGPU/private-element-size.ll b/llvm/test/CodeGen/AMDGPU/private-element-size.ll index 9e75cc2bae4..eb487637ce9 100644 --- a/llvm/test/CodeGen/AMDGPU/private-element-size.ll +++ b/llvm/test/CodeGen/AMDGPU/private-element-size.ll @@ -32,10 +32,10 @@ ; HSA-ELT4-DAG: buffer_store_dword {{v[0-9]+}}, off, s[0:3], s9 offset:40{{$}} ; HSA-ELT4-DAG: buffer_store_dword {{v[0-9]+}}, off, s[0:3], s9 offset:44{{$}} -; HSA-ELT4-DAG: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen{{$}} -; HSA-ELT4-DAG: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:4{{$}} -; HSA-ELT4-DAG: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:8{{$}} -; HSA-ELT4-DAG: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:12{{$}} +; HSA-ELT4: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen{{$}} +; HSA-ELT4: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:4{{$}} +; HSA-ELT4: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:8{{$}} +; HSA-ELT4: buffer_load_dword {{v[0-9]+}}, v{{[0-9]+}}, s[0:3], s9 offen offset:12{{$}} define void @private_elt_size_v4i32(<4 x i32> addrspace(1)* %out, i32 addrspace(1)* %index.array) #0 { entry: %tid = call i32 @llvm.amdgcn.workitem.id.x() @@ -130,8 +130,8 @@ entry: ; HSA-ELT8: private_element_size = 2 ; HSA-ELT4: private_element_size = 1 -; HSA-ELTGE8-DAG: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, {{off|v[0-9]}}, s[0:3], s9 offset:1 -; HSA-ELTGE8-DAG: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, {{off|v[0-9]}}, s[0:3], s9 offset:2 +; HSA-ELTGE8-DAG: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], s9 offset:16 +; HSA-ELTGE8-DAG: buffer_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, off, s[0:3], s9 offset:24 ; HSA-ELTGE8: buffer_load_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, s[0:3], s9 offen diff --git a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll index c1d691fcff8..4beefb047f2 100644 --- a/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll +++ b/llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll @@ -157,8 +157,9 @@ define void @reorder_global_load_local_store_global_load(i32 addrspace(1)* %out, ; FUNC-LABEL: @reorder_local_offsets ; CI: ds_read2_b32 {{v\[[0-9]+:[0-9]+\]}}, {{v[0-9]+}} offset0:100 offset1:102 -; CI-DAG: ds_write2_b32 {{v[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset0:3 offset1:100 -; CI-DAG: ds_write_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:408 +; CI: ds_write2_b32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}} offset0:3 offset1:100 +; CI: ds_read_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:12 +; CI: ds_write_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:408 ; CI: buffer_store_dword ; CI: s_endpgm define void @reorder_local_offsets(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* noalias nocapture readnone %gptr, i32 addrspace(3)* noalias nocapture %ptr0) #0 { @@ -180,12 +181,12 @@ define void @reorder_local_offsets(i32 addrspace(1)* nocapture %out, i32 addrspa } ; FUNC-LABEL: @reorder_global_offsets -; CI-DAG: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400 -; CI-DAG: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:408 -; CI-DAG: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12 -; CI-DAG: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400 -; CI-DAG: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:408 -; CI: buffer_store_dword +; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400 +; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:408 +; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12 +; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:400 +; CI: buffer_store_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:408 +; CI: buffer_load_dword {{v[0-9]+}}, off, {{s\[[0-9]+:[0-9]+\]}}, 0 offset:12 ; CI: s_endpgm define void @reorder_global_offsets(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* noalias nocapture readnone %gptr, i32 addrspace(1)* noalias nocapture %ptr0) #0 { %ptr1 = getelementptr inbounds i32, i32 addrspace(1)* %ptr0, i32 3 |
