maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/AMDGPU/bfe-patterns.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

162 lines
6.9 KiB
LLVM
Raw Normal View History

AMDGPU: Add pass to lower kernel arguments to loads This replaces most argument uses with loads, but for now not all. The code in SelectionDAG for calling convention lowering is actively harmful for amdgpu_kernel. It attempts to split the argument types into register legal types, which results in low quality code for arbitary types. Since all kernel arguments are passed in memory, we just want the raw types. I've tried a couple of methods of mitigating this in SelectionDAG, but it's easier to just bypass this problem alltogether. It's possible to hack around the problem in the initial lowering, but the real problem is the DAG then expects to be able to use CopyToReg/CopyFromReg for uses of the arguments outside the block. Exposing the argument loads in the IR also has the advantage that the LoadStoreVectorizer can merge them. I'm not sure the best approach to dealing with the IR argument list is. The patch as-is just leaves the IR arguments in place, so all the existing code will still compute the same kernarg size and pointlessly lowers the arguments. Arguably the frontend should emit kernels with an empty argument list in the first place. Alternatively a dummy array could be inserted as a single argument just to reserve space. This does have some disadvantages. Local pointer kernel arguments can no longer have AssertZext placed on them as the equivalent !range metadata is not valid on pointer typed loads. This is mostly bad for SI which needs to know about the known bits in order to use the DS instruction offset, so in this case this is not done. More importantly, this skips noalias arguments since this pass does not yet convert this to the equivalent !alias.scope and !noalias metadata. Producing this metadata correctly seems to be tricky, although this logically is the same as inlining into a function which doesn't exist. Additionally, exposing these loads to the vectorizer may result in degraded aliasing information if a pointer load is merged with another argument load. I'm also not entirely sure this is preserving the current clover ABI, although I would greatly prefer if it would stop widening arguments and match the HSA ABI. As-is I think it is extending < 4-byte arguments to 4-bytes but doesn't align them to 4-bytes. llvm-svn: 335650
2018-06-27 03:10:00 +08:00
; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=SI %s
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=VI %s
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
; GCN-LABEL: {{^}}v_ubfe_sub_i32:
; GCN: {{buffer|flat}}_load_dword [[SRC:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[WIDTH:v[0-9]+]]
; GCN: v_bfe_u32 v{{[0-9]+}}, [[SRC]], 0, [[WIDTH]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @v_ubfe_sub_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in0, i32 addrspace(1)* %in1) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in0.gep = getelementptr i32, i32 addrspace(1)* %in0, i32 %id.x
%in1.gep = getelementptr i32, i32 addrspace(1)* %in1, i32 %id.x
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%src = load volatile i32, i32 addrspace(1)* %in0.gep
%width = load volatile i32, i32 addrspace(1)* %in0.gep
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = lshr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
ret void
}
; GCN-LABEL: {{^}}v_ubfe_sub_multi_use_shl_i32:
; GCN: {{buffer|flat}}_load_dword [[SRC:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[WIDTH:v[0-9]+]]
[AMDGPU][MC][GFX8][GFX9] Corrected names of integer v_{add/addc/sub/subrev/subb/subbrev} See bug 34765: https://bugs.llvm.org//show_bug.cgi?id=34765 Reviewers: tamazov, SamWot, arsenm, vpykhtin Differential Revision: https://reviews.llvm.org/D40088 llvm-svn: 318675
2017-11-21 02:24:21 +08:00
; GCN: v_sub_{{[iu]}}32_e32 [[SUB:v[0-9]+]], vcc, 32, [[WIDTH]]
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
[AMDGPU] fixed divergence driven shift operations selection Differential Revision: https://reviews.llvm.org/D73483 Reviewers: rampitec
2020-02-01 01:49:00 +08:00
; SI-NEXT: v_lshl_b32_e32 [[SHL:v[0-9]+]], [[SRC]], [[SUB]]
; SI-NEXT: v_lshr_b32_e32 [[BFE:v[0-9]+]], [[SHL]], [[SUB]]
; VI-NEXT: v_lshlrev_b32_e32 [[SHL:v[0-9]+]], [[SUB]], [[SRC]]
; VI-NEXT: v_lshrrev_b32_e32 [[BFE:v[0-9]+]], [[SUB]], [[SHL]]
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
; GCN: [[BFE]]
; GCN: [[SHL]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @v_ubfe_sub_multi_use_shl_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in0, i32 addrspace(1)* %in1) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in0.gep = getelementptr i32, i32 addrspace(1)* %in0, i32 %id.x
%in1.gep = getelementptr i32, i32 addrspace(1)* %in1, i32 %id.x
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%src = load volatile i32, i32 addrspace(1)* %in0.gep
%width = load volatile i32, i32 addrspace(1)* %in0.gep
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = lshr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
store volatile i32 %shl, i32 addrspace(1)* undef
ret void
}
; GCN-LABEL: {{^}}s_ubfe_sub_i32:
AMDGPU: Add pass to lower kernel arguments to loads This replaces most argument uses with loads, but for now not all. The code in SelectionDAG for calling convention lowering is actively harmful for amdgpu_kernel. It attempts to split the argument types into register legal types, which results in low quality code for arbitary types. Since all kernel arguments are passed in memory, we just want the raw types. I've tried a couple of methods of mitigating this in SelectionDAG, but it's easier to just bypass this problem alltogether. It's possible to hack around the problem in the initial lowering, but the real problem is the DAG then expects to be able to use CopyToReg/CopyFromReg for uses of the arguments outside the block. Exposing the argument loads in the IR also has the advantage that the LoadStoreVectorizer can merge them. I'm not sure the best approach to dealing with the IR argument list is. The patch as-is just leaves the IR arguments in place, so all the existing code will still compute the same kernarg size and pointlessly lowers the arguments. Arguably the frontend should emit kernels with an empty argument list in the first place. Alternatively a dummy array could be inserted as a single argument just to reserve space. This does have some disadvantages. Local pointer kernel arguments can no longer have AssertZext placed on them as the equivalent !range metadata is not valid on pointer typed loads. This is mostly bad for SI which needs to know about the known bits in order to use the DS instruction offset, so in this case this is not done. More importantly, this skips noalias arguments since this pass does not yet convert this to the equivalent !alias.scope and !noalias metadata. Producing this metadata correctly seems to be tricky, although this logically is the same as inlining into a function which doesn't exist. Additionally, exposing these loads to the vectorizer may result in degraded aliasing information if a pointer load is merged with another argument load. I'm also not entirely sure this is preserving the current clover ABI, although I would greatly prefer if it would stop widening arguments and match the HSA ABI. As-is I think it is extending < 4-byte arguments to 4-bytes but doesn't align them to 4-bytes. llvm-svn: 335650
2018-06-27 03:10:00 +08:00
; GCN: s_load_dwordx2 s{{\[}}[[SRC:[0-9]+]]:[[WIDTH:[0-9]+]]{{\]}}, s[0:1], {{0xb|0x2c}}
[AMDGPU] Make bfe patterns divergence-aware This tends to increase code size but more importantly it reduces vgpr usage, and could avoid costly readfirstlanes if the result needs to be in an sgpr. Differential Revision: https://reviews.llvm.org/D88580
2020-09-25 23:07:27 +08:00
; GCN: s_sub_i32 [[SUB:s[0-9]+]], 32, s[[WIDTH]]
; GCN: s_lshl_b32 [[TMP:s[0-9]+]], s[[SRC]], [[SUB]]
; GCN: s_lshr_b32 s{{[0-9]+}}, [[TMP]], [[SUB]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @s_ubfe_sub_i32(i32 addrspace(1)* %out, i32 %src, i32 %width) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = lshr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
ret void
}
; GCN-LABEL: {{^}}s_ubfe_sub_multi_use_shl_i32:
AMDGPU: Add pass to lower kernel arguments to loads This replaces most argument uses with loads, but for now not all. The code in SelectionDAG for calling convention lowering is actively harmful for amdgpu_kernel. It attempts to split the argument types into register legal types, which results in low quality code for arbitary types. Since all kernel arguments are passed in memory, we just want the raw types. I've tried a couple of methods of mitigating this in SelectionDAG, but it's easier to just bypass this problem alltogether. It's possible to hack around the problem in the initial lowering, but the real problem is the DAG then expects to be able to use CopyToReg/CopyFromReg for uses of the arguments outside the block. Exposing the argument loads in the IR also has the advantage that the LoadStoreVectorizer can merge them. I'm not sure the best approach to dealing with the IR argument list is. The patch as-is just leaves the IR arguments in place, so all the existing code will still compute the same kernarg size and pointlessly lowers the arguments. Arguably the frontend should emit kernels with an empty argument list in the first place. Alternatively a dummy array could be inserted as a single argument just to reserve space. This does have some disadvantages. Local pointer kernel arguments can no longer have AssertZext placed on them as the equivalent !range metadata is not valid on pointer typed loads. This is mostly bad for SI which needs to know about the known bits in order to use the DS instruction offset, so in this case this is not done. More importantly, this skips noalias arguments since this pass does not yet convert this to the equivalent !alias.scope and !noalias metadata. Producing this metadata correctly seems to be tricky, although this logically is the same as inlining into a function which doesn't exist. Additionally, exposing these loads to the vectorizer may result in degraded aliasing information if a pointer load is merged with another argument load. I'm also not entirely sure this is preserving the current clover ABI, although I would greatly prefer if it would stop widening arguments and match the HSA ABI. As-is I think it is extending < 4-byte arguments to 4-bytes but doesn't align them to 4-bytes. llvm-svn: 335650
2018-06-27 03:10:00 +08:00
; GCN: s_load_dwordx2 s{{\[}}[[SRC:[0-9]+]]:[[WIDTH:[0-9]+]]{{\]}}, s[0:1], {{0xb|0x2c}}
; GCN: s_sub_i32 [[SUB:s[0-9]+]], 32, s[[WIDTH]]
; GCN: s_lshl_b32 [[SHL:s[0-9]+]], s[[SRC]], [[SUB]]
; GCN: s_lshr_b32 s{{[0-9]+}}, [[SHL]], [[SUB]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @s_ubfe_sub_multi_use_shl_i32(i32 addrspace(1)* %out, i32 %src, i32 %width) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = lshr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
store volatile i32 %shl, i32 addrspace(1)* undef
ret void
}
; GCN-LABEL: {{^}}v_sbfe_sub_i32:
; GCN: {{buffer|flat}}_load_dword [[SRC:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[WIDTH:v[0-9]+]]
; GCN: v_bfe_i32 v{{[0-9]+}}, [[SRC]], 0, [[WIDTH]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @v_sbfe_sub_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in0, i32 addrspace(1)* %in1) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in0.gep = getelementptr i32, i32 addrspace(1)* %in0, i32 %id.x
%in1.gep = getelementptr i32, i32 addrspace(1)* %in1, i32 %id.x
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%src = load volatile i32, i32 addrspace(1)* %in0.gep
%width = load volatile i32, i32 addrspace(1)* %in0.gep
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = ashr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
ret void
}
; GCN-LABEL: {{^}}v_sbfe_sub_multi_use_shl_i32:
; GCN: {{buffer|flat}}_load_dword [[SRC:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[WIDTH:v[0-9]+]]
[AMDGPU][MC][GFX8][GFX9] Corrected names of integer v_{add/addc/sub/subrev/subb/subbrev} See bug 34765: https://bugs.llvm.org//show_bug.cgi?id=34765 Reviewers: tamazov, SamWot, arsenm, vpykhtin Differential Revision: https://reviews.llvm.org/D40088 llvm-svn: 318675
2017-11-21 02:24:21 +08:00
; GCN: v_sub_{{[iu]}}32_e32 [[SUB:v[0-9]+]], vcc, 32, [[WIDTH]]
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
[AMDGPU] fixed divergence driven shift operations selection Differential Revision: https://reviews.llvm.org/D73483 Reviewers: rampitec
2020-02-01 01:49:00 +08:00
; SI-NEXT: v_lshl_b32_e32 [[SHL:v[0-9]+]], [[SRC]], [[SUB]]
; SI-NEXT: v_ashr_i32_e32 [[BFE:v[0-9]+]], [[SHL]], [[SUB]]
; VI-NEXT: v_lshlrev_b32_e32 [[SHL:v[0-9]+]], [[SUB]], [[SRC]]
; VI-NEXT: v_ashrrev_i32_e32 [[BFE:v[0-9]+]], [[SUB]], [[SHL]]
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
; GCN: [[BFE]]
; GCN: [[SHL]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @v_sbfe_sub_multi_use_shl_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in0, i32 addrspace(1)* %in1) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%in0.gep = getelementptr i32, i32 addrspace(1)* %in0, i32 %id.x
%in1.gep = getelementptr i32, i32 addrspace(1)* %in1, i32 %id.x
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%src = load volatile i32, i32 addrspace(1)* %in0.gep
%width = load volatile i32, i32 addrspace(1)* %in0.gep
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = ashr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
store volatile i32 %shl, i32 addrspace(1)* undef
ret void
}
; GCN-LABEL: {{^}}s_sbfe_sub_i32:
AMDGPU: Add pass to lower kernel arguments to loads This replaces most argument uses with loads, but for now not all. The code in SelectionDAG for calling convention lowering is actively harmful for amdgpu_kernel. It attempts to split the argument types into register legal types, which results in low quality code for arbitary types. Since all kernel arguments are passed in memory, we just want the raw types. I've tried a couple of methods of mitigating this in SelectionDAG, but it's easier to just bypass this problem alltogether. It's possible to hack around the problem in the initial lowering, but the real problem is the DAG then expects to be able to use CopyToReg/CopyFromReg for uses of the arguments outside the block. Exposing the argument loads in the IR also has the advantage that the LoadStoreVectorizer can merge them. I'm not sure the best approach to dealing with the IR argument list is. The patch as-is just leaves the IR arguments in place, so all the existing code will still compute the same kernarg size and pointlessly lowers the arguments. Arguably the frontend should emit kernels with an empty argument list in the first place. Alternatively a dummy array could be inserted as a single argument just to reserve space. This does have some disadvantages. Local pointer kernel arguments can no longer have AssertZext placed on them as the equivalent !range metadata is not valid on pointer typed loads. This is mostly bad for SI which needs to know about the known bits in order to use the DS instruction offset, so in this case this is not done. More importantly, this skips noalias arguments since this pass does not yet convert this to the equivalent !alias.scope and !noalias metadata. Producing this metadata correctly seems to be tricky, although this logically is the same as inlining into a function which doesn't exist. Additionally, exposing these loads to the vectorizer may result in degraded aliasing information if a pointer load is merged with another argument load. I'm also not entirely sure this is preserving the current clover ABI, although I would greatly prefer if it would stop widening arguments and match the HSA ABI. As-is I think it is extending < 4-byte arguments to 4-bytes but doesn't align them to 4-bytes. llvm-svn: 335650
2018-06-27 03:10:00 +08:00
; GCN: s_load_dwordx2 s{{\[}}[[SRC:[0-9]+]]:[[WIDTH:[0-9]+]]{{\]}}, s[0:1], {{0xb|0x2c}}
[AMDGPU] Make bfe patterns divergence-aware This tends to increase code size but more importantly it reduces vgpr usage, and could avoid costly readfirstlanes if the result needs to be in an sgpr. Differential Revision: https://reviews.llvm.org/D88580
2020-09-25 23:07:27 +08:00
; GCN: s_sub_i32 [[SUB:s[0-9]+]], 32, s[[WIDTH]]
; GCN: s_lshl_b32 [[TMP:s[0-9]+]], s[[SRC]], [[SUB]]
; GCN: s_ashr_i32 s{{[0-9]+}}, [[TMP]], [[SUB]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @s_sbfe_sub_i32(i32 addrspace(1)* %out, i32 %src, i32 %width) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = ashr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
ret void
}
; GCN-LABEL: {{^}}s_sbfe_sub_multi_use_shl_i32:
AMDGPU: Add pass to lower kernel arguments to loads This replaces most argument uses with loads, but for now not all. The code in SelectionDAG for calling convention lowering is actively harmful for amdgpu_kernel. It attempts to split the argument types into register legal types, which results in low quality code for arbitary types. Since all kernel arguments are passed in memory, we just want the raw types. I've tried a couple of methods of mitigating this in SelectionDAG, but it's easier to just bypass this problem alltogether. It's possible to hack around the problem in the initial lowering, but the real problem is the DAG then expects to be able to use CopyToReg/CopyFromReg for uses of the arguments outside the block. Exposing the argument loads in the IR also has the advantage that the LoadStoreVectorizer can merge them. I'm not sure the best approach to dealing with the IR argument list is. The patch as-is just leaves the IR arguments in place, so all the existing code will still compute the same kernarg size and pointlessly lowers the arguments. Arguably the frontend should emit kernels with an empty argument list in the first place. Alternatively a dummy array could be inserted as a single argument just to reserve space. This does have some disadvantages. Local pointer kernel arguments can no longer have AssertZext placed on them as the equivalent !range metadata is not valid on pointer typed loads. This is mostly bad for SI which needs to know about the known bits in order to use the DS instruction offset, so in this case this is not done. More importantly, this skips noalias arguments since this pass does not yet convert this to the equivalent !alias.scope and !noalias metadata. Producing this metadata correctly seems to be tricky, although this logically is the same as inlining into a function which doesn't exist. Additionally, exposing these loads to the vectorizer may result in degraded aliasing information if a pointer load is merged with another argument load. I'm also not entirely sure this is preserving the current clover ABI, although I would greatly prefer if it would stop widening arguments and match the HSA ABI. As-is I think it is extending < 4-byte arguments to 4-bytes but doesn't align them to 4-bytes. llvm-svn: 335650
2018-06-27 03:10:00 +08:00
; GCN: s_load_dwordx2 s{{\[}}[[SRC:[0-9]+]]:[[WIDTH:[0-9]+]]{{\]}}, s[0:1], {{0xb|0x2c}}
; GCN: s_sub_i32 [[SUB:s[0-9]+]], 32, s[[WIDTH]]
; GCN: s_lshl_b32 [[SHL:s[0-9]+]], s[[SRC]], [[SUB]]
; GCN: s_ashr_i32 s{{[0-9]+}}, [[SHL]], [[SUB]]
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
2017-03-22 05:39:51 +08:00
define amdgpu_kernel void @s_sbfe_sub_multi_use_shl_i32(i32 addrspace(1)* %out, i32 %src, i32 %width) #1 {
AMDGPU: Add another BFE pattern This is the pattern that falls out of the instruction's definition if offset == 0. llvm-svn: 295912
2017-02-23 08:23:43 +08:00
%id.x = tail call i32 @llvm.amdgcn.workitem.id.x()
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id.x
%sub = sub i32 32, %width
%shl = shl i32 %src, %sub
%bfe = ashr i32 %shl, %sub
store i32 %bfe, i32 addrspace(1)* %out.gep
store volatile i32 %shl, i32 addrspace(1)* undef
ret void
}
declare i32 @llvm.amdgcn.workitem.id.x() #0
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind }