[PATCH] D145401: [AMDGPU] Reserve extra SGPR blocks wth XNACK "any" TID Setting

This revision was automatically updated to reflect the committed changes. Closed by commit rG864a2b25beac: [AMDGPU] Reserve extra SGPR blocks wth XNACK any TID Setting (authored by kerbowa). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145401/new/

[PATCH] D145401: [AMDGPU] Reserve extra SGPR blocks wth XNACK "any" TID Setting

kerbowa added a comment. Added AMDGPU group to reviewers. Is there any objection to changing the defaults for subtargets that support XNACK to always reserve extra SGPRs unless -xnack is explicitly requested? This would impact graphics as well. The old defaults were doing the opposite and

[PATCH] D145401: [AMDGPU] Reserve extra SGPR blocks wth XNACK "any" TID Setting

kerbowa updated this revision to Diff 503461. kerbowa added a comment. Herald added a project: clang. Herald added a subscriber: cfe-commits. Update tests. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D145401/new/ https://reviews.llvm.org/D145401

[PATCH] D132079: [AMDGPU] Add iglp_opt builtin and MFMA GEMM Opt strategy

kerbowa added inline comments. Comment at: llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp:427 DAG->addMutation(createStoreClusterDAGMutation(DAG->TII, DAG->TRI)); DAG->addMutation(createIGroupLPDAGMutation()); DAG->addMutation(createAMDGPUMacroFusionDAGMutation());

[PATCH] D128158: [AMDGPU] Add amdgcn_sched_group_barrier builtin

kerbowa marked an inline comment as done. kerbowa added inline comments. Comment at: llvm/lib/Target/AMDGPU/AMDGPUIGroupLP.cpp:314 + +bool SchedGroup::isFull() const { + return MaxSize && Collection.size() >= *MaxSize; uabelho wrote: > Compiling with gcc, I get

[PATCH] D128158: [AMDGPU] Add amdgcn_sched_group_barrier builtin

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGf5b21680d122: [AMDGPU] Add amdgcn_sched_group_barrier builtin (authored by kerbowa). Changed prior to commit:

[PATCH] D128158: [AMDGPU] Add amdgcn_sched_group_barrier builtin

kerbowa updated this revision to Diff 445965. kerbowa added a comment. Fix some bugs. Add better pipeline fitting. Address comments. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D128158/new/ https://reviews.llvm.org/D128158 Files:

[PATCH] D128158: [AMDGPU] Add amdgcn_sched_group_barrier builtin

kerbowa added a comment. Somewhat WIP needs more tests and cleanup. Posted for dependent work. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D128158/new/ https://reviews.llvm.org/D128158 ___ cfe-commits

[PATCH] D128158: [AMDGPU] Add amdgcn_sched_group_barrier builtin

kerbowa created this revision. kerbowa added reviewers: rampitec, jrbyrnes, vangthao95, arsenm. Herald added subscribers: kosarev, jsilvanus, foad, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, jvesely, kzhuravl. Herald added a project: All. kerbowa requested review of this revision. Herald

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

This revision was automatically updated to reflect the committed changes. Closed by commit rG2db700215a2e: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic (authored by kerbowa). Herald added a subscriber: kosarev. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

kerbowa updated this revision to Diff 427747. kerbowa added a comment. Herald added a subscriber: jsilvanus. Use i32. Output hex. Fix hazard rec tests for pseudo instructions. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D124700/new/

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

kerbowa added a comment. In D124700#3483633 , @rampitec wrote: > In D124700#3483609 , @kerbowa wrote: > >> In D124700#3483556 , @rampitec >> wrote: >> >>> You do not

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

kerbowa added a comment. In D124700#3483556 , @rampitec wrote: > You do not handle masks other than 0 yet? We handle 0 and 1 only. Comment at: llvm/include/llvm/IR/IntrinsicsAMDGPU.td:219 +// MASK = 0: No instructions may be

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

kerbowa updated this revision to Diff 426169. kerbowa added a comment. Add mir tests. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D124700/new/ https://reviews.llvm.org/D124700 Files: clang/include/clang/Basic/BuiltinsAMDGPU.def

[PATCH] D124700: [AMDGPU] Add llvm.amdgcn.sched.barrier intrinsic

kerbowa created this revision. Herald added subscribers: hsmhsm, foad, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, jvesely, kzhuravl, arsenm. Herald added a project: All. kerbowa requested review of this revision. Herald added subscribers: llvm-commits, cfe-commits, wdng. Herald added

[PATCH] D120976: [AMDGPU] Add llvm.amdgcn.s.setprio intrinsic

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG62bcfcb5a588: [AMDGPU] Add llvm.amdgcn.s.setprio intrinsic (authored by kerbowa). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D120976: [AMDGPU] Add llvm.amdgcn.s.setprio intrinsic

kerbowa updated this revision to Diff 414559. kerbowa added a comment. Herald added a project: clang. Herald added a subscriber: cfe-commits. Add clang builtin and tests. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D120976/new/

[PATCH] D79213: [hip] Add noalias on restrict qualified coerced hip pointers

kerbowa added a comment. In D79213#2018820 , @hliao wrote: > Any more comments? As this should be a performance-critical issue, shall we > get conclusion and make progress for the next step? We applied this current version of the patch internally for

[PATCH] D79213: [hip] Add noalias on restrict qualified coerced hip pointers

kerbowa added inline comments. Comment at: clang/lib/CodeGen/CGCall.cpp:2270 + CGF.ConvertType(Ty)->getPointerAddressSpace() == 0 && + ArgI.getCoerceToType()->getPointerElementType() == + CGF.ConvertType(Ty)->getPointerElementType();

[PATCH] D79213: [hip] Add noalias on restrict qualified coerced hip pointers

kerbowa added inline comments. Comment at: clang/lib/CodeGen/CGCall.cpp:2270 + CGF.ConvertType(Ty)->getPointerAddressSpace() == 0 && + ArgI.getCoerceToType()->getPointerElementType() == + CGF.ConvertType(Ty)->getPointerElementType();

[PATCH] D79213: [hip] Add noalias on restrict qualified coerced hip pointers

kerbowa updated this revision to Diff 261388. kerbowa added a comment. Fix test formatting. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D79213/new/ https://reviews.llvm.org/D79213 Files: clang/lib/CodeGen/CGCall.cpp

[PATCH] D79213: [hip] Add noalias on restrict qualified coerced hip pointers

kerbowa created this revision. kerbowa added reviewers: yaxunl, hliao. Herald added subscribers: cfe-commits, nhaehnle, jvesely. Herald added a project: clang. Repository: rG LLVM Github Monorepo https://reviews.llvm.org/D79213 Files: clang/lib/CodeGen/CGCall.cpp