[PATCH] D134824: Fix frint ACLE intrinsic names

jaykang10 accepted this revision. jaykang10 added a comment. This revision is now accepted and ready to land. Thanks for fixing it! LGTM Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D134824/new/ https://reviews.llvm.org/D134824

[PATCH] D108470: [OpenCL] Fix as_type3 invalid store creation

jaykang10 added a comment. In D108470#2959715 , @svenvh wrote: > In D108470#2959708 , @jaykang10 > wrote: > >> Can you also add "double2_to_float3" > > Instead of `double2_to_float3`, I decided to add

[PATCH] D108470: [OpenCL] Fix as_type3 invalid store creation

jaykang10 added a comment. Can you also add "double2_to_float3" and "float4_to_float3" tests please? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108470/new/ https://reviews.llvm.org/D108470 ___ cfe-commits mailing list

[PATCH] D108470: [OpenCL] Fix as_type3 invalid store creation

jaykang10 added a comment. If possible, can you add more tests for different types as previous patch please? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108470/new/ https://reviews.llvm.org/D108470

[PATCH] D107963: [OpenCL] Fix as_type(vec3) invalid store creation

jaykang10 added inline comments. Comment at: clang/lib/CodeGen/CGExprScalar.cpp:4789 - -if (!CGF.CGM.getCodeGenOpts().PreserveVec3Type) { - Src = createCastsForTypeOfSameSize(Builder, CGF.CGM.getDataLayout(), Src, svenvh wrote: > jaykang10 wrote: > >

[PATCH] D107963: [OpenCL] Fix as_type(vec3) invalid store creation

jaykang10 added inline comments. Comment at: clang/lib/CodeGen/CGExprScalar.cpp:4789 - -if (!CGF.CGM.getCodeGenOpts().PreserveVec3Type) { - Src = createCastsForTypeOfSameSize(Builder, CGF.CGM.getDataLayout(), Src, svenvh wrote: > Anastasia wrote: > >

[PATCH] D98269: [AArch64] Add missing intrinsics for scalar fp rounding

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG25951c5ab8e9: [AArch64] Add missing intrinsics for scalar FP rounding (authored by jaykang10). Herald added a project: clang. Herald added a

[PATCH] D97775: [AArch64] Add missing intrinsics for vcls

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG394a4d04333e: [AArch64] Add missing intrinsics for vcls (authored by jaykang10). Herald added a project: clang. Herald added a subscriber:

[PATCH] D62962: Clang implementation of sizeless types

jaykang10 added inline comments. Comment at: test/Sema/sizeless-1.c:29 + +typedef svint8_t vec_int8 __attribute__((vector_size(64))); // expected-error {{invalid vector element type}} + Please check "__attribute__((ext_vector_type()))". I guess it is also

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#718176, @Anastasia wrote: > LGTM! Thanks! Thanks Anastasia! Merge Done! Repository: rL LLVM https://reviews.llvm.org/D30810 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D30810: Preserve vec3 type.

This revision was automatically updated to reflect the committed changes. Closed by commit rL299445: Preserve vec3 type. (authored by jaykang10). Changed prior to commit: https://reviews.llvm.org/D30810?vs=92829=94088#toc Repository: rL LLVM https://reviews.llvm.org/D30810 Files:

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. > I believe the argument lacks numbers (or at least you have them but didn't > mention). I didn't hear about performance results, or validation that this > was actually tested for correctness. Small test cases prove a point, but > can't be considered general. > >

[PATCH] D30810: Preserve vec3 type.

jaykang10 updated this revision to Diff 92829. jaykang10 added a comment. Preserved vec3 type on __builtin_astype. https://reviews.llvm.org/D30810 Files: include/clang/Driver/CC1Options.td include/clang/Frontend/CodeGenOptions.def lib/CodeGen/CGExpr.cpp lib/CodeGen/CGExprScalar.cpp

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. > Yes. This would make sense. I am guessing that in vec3->vec4, we will have 3 > loads and 4 stores and in vec4->vec3 we will have 4 loads and 3 stores? It depends on implementation. If you scalarize all vector operations on LLVM IR level before entering llvm's

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#706677, @Anastasia wrote: > In https://reviews.llvm.org/D30810#706676, @Anastasia wrote: > > > In https://reviews.llvm.org/D30810#699428, @Anastasia wrote: > > > > > Would you be able to update ScalarExprEmitter::VisitAsTypeExpr > >

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#702637, @jaykang10 wrote: > In https://reviews.llvm.org/D30810#702614, @Anastasia wrote: > > > In https://reviews.llvm.org/D30810#702443, @bruno wrote: > > > > > > As a result, I think it would be good for clang to have both of > > >

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#702614, @Anastasia wrote: > In https://reviews.llvm.org/D30810#702443, @bruno wrote: > > > > As a result, I think it would be good for clang to have both of features > > > and I would like to stick to the option "-fpresereve-vec3' to

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#702548, @ahatanak wrote: > In https://reviews.llvm.org/D30810#701141, @jaykang10 wrote: > > > In https://reviews.llvm.org/D30810#701132, @ahatanak wrote: > > > > > Actually, it's not a mis-compile. The record layout shows that there

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. > The motivation doesn't seem solid to me, who else is going to benefit from > this flag? You also didn't explain why doing this transformation yourself > (looking through the shuffle) on your downstream pass isn't enough for you. > We generally try to avoid adding

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#701861, @Anastasia wrote: > I don't think there is anything wrong with the generation of vec3->vec4 in > Clang. I believe the motivation for this was the OpenCL spec treating vec3 as > vec4 aligned type (see section 6.1.5: >

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. > I believe the assumption is more practical: most part of upstream llvm > targets only support vectors with even sized number of lanes. And in those > cases you would have to expand to a 4x vector and leave the 4th element as > undef anyway, so it was done in the

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#701132, @ahatanak wrote: > Actually, it's not a mis-compile. The record layout shows that there is a > padding before field f2 and f2 starts at byte 16. So using "store <4 x > float>" doesn't overwrite the field. It depends on

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#700476, @Anastasia wrote: > In https://reviews.llvm.org/D30810#699827, @jaykang10 wrote: > > > In https://reviews.llvm.org/D30810#699695, @bruno wrote: > > > > > Hi JinGu, > > > > > > I just read the discussion on cfe-dev, some

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#699695, @bruno wrote: > Hi JinGu, > > I just read the discussion on cfe-dev, some comments: > > > My colleague and I are implementing a transformation pass between LLVM IR > > and another IR and we want to keep the 3-component vector

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#699428, @Anastasia wrote: > Would you be able to update ScalarExprEmitter::VisitAsTypeExpr implementation > accordingly to make things consistent? Probably, No... the load/store with vec3 generates vec4 temporarily to be aligned

[PATCH] D30810: Preserve vec3 type.

jaykang10 marked an inline comment as done. jaykang10 added a comment. In https://reviews.llvm.org/D30810#699006, @ahatanak wrote: > Could you elaborate on the motivation for this change? > > I was wondering why clang (CodeGen) needed the help of a command line option > to decide whether vec3

[PATCH] D30810: Preserve vec3 type.

jaykang10 added a comment. In https://reviews.llvm.org/D30810#697760, @Anastasia wrote: > Could you please add your test here (probably goes to test/CodeGenOpenCL)? Oops! I am so sorry. I missed it. I have updated it. https://reviews.llvm.org/D30810

[PATCH] D30810: Preserve vec3 type.

jaykang10 updated this revision to Diff 91355. jaykang10 added a comment. Changed help text for option and Added test file. https://reviews.llvm.org/D30810 Files: include/clang/Driver/CC1Options.td include/clang/Frontend/CodeGenOptions.def lib/CodeGen/CGExpr.cpp

[PATCH] D30810: Preserve vec3 type.

jaykang10 updated this revision to Diff 91307. jaykang10 added a comment. Fixed typo. https://reviews.llvm.org/D30810 Files: include/clang/Driver/CC1Options.td include/clang/Frontend/CodeGenOptions.def lib/CodeGen/CGExpr.cpp lib/Frontend/CompilerInvocation.cpp Index:

[PATCH] D30810: Preserve vec3 type.

jaykang10 updated this revision to Diff 91306. jaykang10 added a comment. Added -f prefix to option name. https://reviews.llvm.org/D30810 Files: include/clang/Driver/CC1Options.td include/clang/Frontend/CodeGenOptions.def lib/CodeGen/CGExpr.cpp lib/Frontend/CompilerInvocation.cpp

[PATCH] D30810: Preserve vec3 type.

jaykang10 created this revision. Preserve vec3 type with CodeGen option. https://reviews.llvm.org/D30810 Files: include/clang/Driver/CC1Options.td include/clang/Frontend/CodeGenOptions.def lib/CodeGen/CGExpr.cpp lib/Frontend/CompilerInvocation.cpp test/CodeGenOpenCL/preserve_vec3.cl