[PATCH] D91157: [AArch64] Out-of-line atomics (-moutline-atomics) implementation.

sebpop added a comment. Herald added a subscriber: MaskRay. Herald added a project: All. Hi Pavel, We need to handle one more case for __sync_* builtins, please see testcase and patches applied to GCC: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105162 Repository: rG LLVM Github Monorepo C

[PATCH] D91157: [AArch64] Out-of-line atomics (-moutline-atomics) implementation.

sebpop added a comment. I tested this change on Graviton2 aarch64-linux by building https://github.com/xianyi/OpenBLAS with `clang -O3 -moutline-atomics` and `make test`: all tests pass with and without outline-atomics. Clang was configured to use libgcc. I also tested https://github.com/boosto

[PATCH] D67368: [NFCI]Create CommonAttributeInfo Type as base type of *Attr and ParsedAttr.

sebpop added a comment. I still see a link error on aarch64-linux on master: /usr/bin/ld: tools/clang/lib/AST/CMakeFiles/obj.clangAST.dir/AttrImpl.cpp.o: in function `clang::AttributeCommonInfo::getAttributeSpellingListIndex() const': /home/ubuntu/llvm-project/llvm/tools/clang/include/clang

[PATCH] D30035: Add const to function parameters

sebpop added a comment. Looks good to me. https://reviews.llvm.org/D30035 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D27068: Improve string::find

sebpop added a comment. In https://reviews.llvm.org/D27068#632685, @EricWF wrote: > Holy crap those improvements are impressive. You're welcome. > This LGTM. I'm assuming @mclow.lists has nothing left to say about this. Thanks for your review and LGTM. I addressed your last comment, added som

[PATCH] D27068: Improve string::find

This revision was automatically updated to reflect the committed changes. Closed by commit rL290761: improve performance of string::find (authored by spop). Changed prior to commit: https://reviews.llvm.org/D27068?vs=80134&id=82731#toc Repository: rL LLVM https://reviews.llvm.org/D27068 Fi

[PATCH] D24991: Inline hot functions in libcxx shared_ptr implementation.

sebpop added a comment. @mclow.lists could you please have a last look at this patch: the change is for a performance improvement (20% uplift on a proprietary benchmark), and all the issues mentioned in the review have been addressed. The existing synthetic benchmark shows an overall improvement

[PATCH] D27068: Improve string::find

sebpop added a comment. The numbers are from an x86_64-linux machine: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz Overall the patch is a win on the synthetic benchmark. We have also seen this in a proprietary benchmark where it accounted for about 10% speedup. https://reviews.llvm.org/D27068 _

[PATCH] D27068: Improve string::find

sebpop added a comment. In https://reviews.llvm.org/D27068#606757, @mclow.lists wrote: > Definitely want to see numbers. master: Benchmark Time CPU Iterations --- BM_StringFindPhase1/10

[PATCH] D26991: Hoist redundant load

sebpop added a comment. In https://reviews.llvm.org/D26991#606764, @mclow.lists wrote: > /me wonders what the perf difference when `_LIBCPP_UNROLL_LOOPS` is defined > or not. > > I think this (`_LIBCPP_UNROLL_LOOPS`) falls squarely into Chandler's request > that we complain to him when the com

[PATCH] D27068: Improve string::find

sebpop added a comment. Please also post the performance numbers from the added benchmarks with and without the change to the lib. Comment at: libcxx/benchmarks/string.bench.cpp:12 +// until the end of s1. +static void BM_StringFindPhase1(benchmark::State& state) { + // Bench

[PATCH] D26991: Hoist redundant load

sebpop added a comment. Let's also add a testcase and show the performance improvement. https://reviews.llvm.org/D26991 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits