aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/CodeGen/ExpandMemCmp.cpp
AgeCommit message (Collapse)AuthorFilesLines
2019-06-26Revert "r364412 [ExpandMemCmp][MergeICmps] Move passes out of CodeGen into ↵Clement Courbet1-0/+871
opt pipeline." Breaks sanitizers: libFuzzer :: cxxstring.test libFuzzer :: memcmp.test libFuzzer :: recommended-dictionary.test libFuzzer :: strcmp.test libFuzzer :: value-profile-mem.test libFuzzer :: value-profile-strcmp.test llvm-svn: 364416
2019-06-26[ExpandMemCmp][MergeICmps] Move passes out of CodeGen into opt pipeline.Clement Courbet1-871/+0
This allows later passes (in particular InstCombine) to optimize more cases. One that's important to us is `memcmp(p, q, constant) < 0` and memcmp(p, q, constant) > 0. llvm-svn: 364412
2019-06-25[ExpandMemCmp] Move all options to TargetTransformInfo.Clement Courbet1-29/+20
Split off from D60318. llvm-svn: 364281
2019-05-18MemCmpExpansion::getCompareLoadPairs - assert we find a comparison diff. NFCI.Simon Pilgrim1-1/+3
Fix scan-build uninitialized warning and assert the final diff isn't null. llvm-svn: 361095
2019-04-12Add options for MaxLoadsPerMemcmp(OptSize).Hiroshi Yamauchi1-2/+15
Reviewers: davidxl Reviewed By: davidxl Subscribers: llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D60587 llvm-svn: 358287
2019-04-04[IR] Refactor attribute methods in Function class (NFC)Evandro Menezes1-2/+2
Rename the functions that query the optimization kind attributes. Differential revision: https://reviews.llvm.org/D60287 llvm-svn: 357731
2019-03-20[ExpandMemCmp] Trigger on bcmp too.Clement Courbet1-1/+2
Summary: Fixes 41150. Reviewers: gchatelet Subscribers: hiraditya, llvm-commits, ckennelly, sbenza, jyknight Tags: #llvm Differential Revision: https://reviews.llvm.org/D59593 llvm-svn: 356550
2019-01-19Update the file headers across all of the LLVM projects in the monorepoChandler Carruth1-4/+3
to reflect the new license. We understand that people may be surprised that we're moving the header entirely to discuss the new license. We checked this carefully with the Foundation's lawyer and we believe this is the correct approach. Essentially, all code in the project is now made available by the LLVM project under our new license, so you will see that the license headers include that license only. Some of our contributors have contributed code under our old license, and accordingly, we have retained a copy of our old license notice in the top-level files in each project and repository. llvm-svn: 351636
2018-12-20Re-land r349731 "[CodeGen][ExpandMemcmp] Add an option for allowing ↵Clement Courbet1-96/+137
overlapping loads. Update PPC ir following GEP->bitcat to bitcat->GEP->bitcat change. llvm-svn: 349747
2018-12-20Revert r349731 "[CodeGen][ExpandMemcmp] Add an option for allowing ↵Clement Courbet1-137/+96
overlapping loads." Forgot to update PowerPC tests for the GEP->bitcast change. llvm-svn: 349733
2018-12-20[CodeGen][ExpandMemcmp] Add an option for allowing overlapping loads.Clement Courbet1-96/+137
Summary: This allows expanding {7,11,13,14,15,21,22,23,25,26,27,28,29,30,31}-byte memcmp in just two loads on X86. These were previously calling memcmp. Reviewers: spatel, gchatelet Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D55263 llvm-svn: 349731
2018-07-16[CodeGen] Fix inconsistent declaration parameter nameFangrui Song1-1/+1
llvm-svn: 337200
2018-01-06[x86, MemCmpExpansion] allow 2 pairs of loads per block (PR33325)Sanjay Patel1-6/+2
This is the last step needed to fix PR33325: https://bugs.llvm.org/show_bug.cgi?id=33325 We're trading branch and compares for loads and logic ops. This makes the code smaller and hopefully faster in most cases. The 24-byte test shows an interesting construct: we load the trailing scalar elements into vector registers and generate the same pcmpeq+movmsk code that we expected for a pair of full vector elements (see the 32- and 64-byte tests). Differential Revision: https://reviews.llvm.org/D41714 llvm-svn: 321934
2018-01-03[ExpandMemcmp] rename variables and add hook to override pref for number of ↵Sanjay Patel1-14/+17
loads per block; NFC The preference only applies to 'memcmp() == 0' expansion, so try to make that clearer. x86 will likely benefit by increasing the default value from '1' to '2' as seen in PR33325: https://bugs.llvm.org/show_bug.cgi?id=33325 ...so that is the planned follow-up to this clean-up step. llvm-svn: 321756
2017-12-18[NFC][CodeGen][ExpandMemCmp] Fix documentation.Clement Courbet1-3/+2
llvm-svn: 320960
2017-12-13Remove redundant includes from lib/CodeGen.Michael Zolotukhin1-2/+0
llvm-svn: 320619
2017-11-17Fix a bunch more layering of CodeGen headers that are in TargetDavid Blaikie1-2/+2
All these headers already depend on CodeGen headers so moving them into CodeGen fixes the layering (since CodeGen depends on Target, not the other way around). llvm-svn: 318490
2017-11-03re-land [ExpandMemCmp] Split ExpandMemCmp from CodeGen into its own pass."Clement Courbet1-0/+828
Fix undefined references: ExpandMemCmp belongs to CodeGen/, not Scalar/. llvm-svn: 317318
generated by cgit v1.1 at 2025-10-09 15:56:15 +0000