Commits · d12c189e2e9e7829220c2836ed38908d7818f2db · Roger Ferrer / llvm-epi · GitLab

Oct 06, 2017
- Revert "[MergeICmps] Disable mergeicmps if the target does not want to handle memcmp expansion." · d12c189e
  Clement Courbet authored Oct 06, 2017
  d12c189e
- Re-land "[MergeICmps] Disable mergeicmps if the target does not want to handle memcmp expansion." · 4e1bae81
  Clement Courbet authored Oct 06, 2017
  4e1bae81
- Roll forward r314928 · bcd36f7c
  Xinliang David Li authored Oct 06, 2017
  bcd36f7c
Oct 05, 2017
- ModuleUtils: Stop using comdat members to generate unique module ids. · 715bcfe0
  Peter Collingbourne authored Oct 05, 2017
  715bcfe0
- [InstCombine] improve folds for icmp gt/lt (shr X, C1), C2 · 7ac2db6a
  Sanjay Patel authored Oct 05, 2017
  7ac2db6a
- Annotate VP prof on indirect call if it is ICPed in the profiled binary. · 16f01fb1
  Dehao Chen authored Oct 05, 2017
  16f01fb1
- [LV] Fix PR34743 - handle casts that sink after interleaved loads · c9e0f886
  Ayal Zaks authored Oct 05, 2017
  c9e0f886
- Revert "Re-land "[MergeICmps] Disable mergeicmps if the target does not want... · 922e5bc6
  Clement Courbet authored Oct 05, 2017
  922e5bc6
- revert r314698 - [InstCombine] remove one-use restriction for icmp (shr exact... · f11b5b4f
  Sanjay Patel authored Oct 05, 2017
  f11b5b4f
- [LV] Fix PR34711 - widen instruction ranges when sinking casts · fc3f7a4f
  Ayal Zaks authored Oct 05, 2017
  fc3f7a4f
- Re-land "[MergeICmps] Disable mergeicmps if the target does not want to handle memcmp expansion."" · 4cafbb9b
  Clement Courbet authored Oct 05, 2017
  4cafbb9b
- Revert "[MergeICmps] Disable mergeicmps if the target does not want to handle memcmp expansion." · 6603fc0e
  Clement Courbet authored Oct 05, 2017
  6603fc0e
- [InstCombine] Fix a vector splat handling bug in transformZExtICmp. · 17b0c784
  Craig Topper authored Oct 05, 2017
  17b0c784
- [MergeICmps] Disable mergeicmps if the target does not want to handle memcmp expansion. · 902eef32
  Clement Courbet authored Oct 05, 2017
  902eef32
- Revert r314928 to investigate thinLTO bootstrap failure · 04ab11a0
  Xinliang David Li authored Oct 05, 2017
  04ab11a0
- [InstCombine] Improve support for ashr in foldICmpAndShift · 7a930923
  Craig Topper authored Oct 04, 2017
  7a930923
Oct 04, 2017
- Recommit r314561 after fixing msan build failure · 7a737573
  Xinliang David Li authored Oct 04, 2017
  7a737573
- Recommit : Use the basic cost if a GEP is not used as addressing mode · d40e03c2
  Jun Bum Lim authored Oct 04, 2017
  d40e03c2
- Revert r314886 "[X86] Improvement in CodeGen instruction selection for LEAs... · 2a6c9adb
  Hans Wennborg authored Oct 04, 2017
  2a6c9adb
- [InstCombine] add 'exact' variants of all tests; NFC · 9366b9c5
  Sanjay Patel authored Oct 04, 2017
  9366b9c5
- [X86] Improvement in CodeGen instruction selection for LEAs (re-applying post... · 3c29bacd
  Jatin Bhateja authored Oct 04, 2017
  3c29bacd
- [IRCE] Temporarily disable unsigned latch conditions by default · 8aacef6c
  Max Kazantsev authored Oct 04, 2017
  8aacef6c
- Move verbosity check for remarks to the diag handler · f31b1f31
  Adam Nemet authored Oct 04, 2017
  f31b1f31
Oct 03, 2017
- [InstCombine] add tests for icmp gt/lt (shr X, C1), C2; NFC · 389b7ced
  Sanjay Patel authored Oct 03, 2017
  389b7ced
- Revert r314806 "[SLP] Vectorize jumbled memory loads." · 9a9048e1
  Hans Wennborg authored Oct 03, 2017
  9a9048e1
- [SLP] Vectorize jumbled memory loads. · 1d5422f2
  Mohammad Shahid authored Oct 03, 2017
  1d5422f2
- [CGP] In optimizeMemoryInst handle select similarly to phi · eb83c755
  John Brawn authored Oct 03, 2017
  eb83c755
- [InstSimplify] teach SimplifySelectInst() to fold more vector selects · 25f6c196
  Haicheng Wu authored Oct 02, 2017
  25f6c196
Oct 02, 2017
- [InstCombine] auto-generate complete checks; NFC · 4eb04617
  Sanjay Patel authored Oct 02, 2017
  4eb04617
- [InstCombine] add icmp (shr X, Y), 0 test; NFC · 5ad57361
  Sanjay Patel authored Oct 02, 2017
  5ad57361
- [InstCombine] remove one-use restriction for icmp (shr exact X, C1), C2 --> icmp X, (C2<<C1) · 6e17c00a
  Sanjay Patel authored Oct 02, 2017
  6e17c00a
- [InstCombine] add icmp (lshr X, C1), C2 test; NFC · ce86ab3e
  Sanjay Patel authored Oct 02, 2017
  ce86ab3e
- Update getMergedLocation to check the instruction type and merge properly. · f464627f
  Dehao Chen authored Oct 02, 2017
  f464627f
- [X86][LLVM]Expanding Supports lowerInterleaved{store|load}() in... · e4084f6b
  Michael Zuckerman authored Oct 02, 2017
  e4084f6b
Oct 01, 2017
- Revert r314579: "Recommi r314561 after fixing over-debug assertion". · 3c9c60c7
  Daniel Jasper authored Oct 01, 2017
  3c9c60c7
- Adding test for interleved, case stride 4 vf64 store<NFC>. · 17468954
  Michael Zuckerman authored Oct 01, 2017
  17468954
- Separate the logic when handling indirect calls in SamplePGO ThinLTO compile... · d26dae0d
  Dehao Chen authored Oct 01, 2017
  d26dae0d
- NewGVN: Fix PR 34473, by not using ExactlyEqualsExpression for finding · d36c27be
  Daniel Berlin authored Sep 30, 2017
  d36c27be
- NewGVN: Evaluate phi of ops expressions before creating phi node · c1305af0
  Daniel Berlin authored Sep 30, 2017
  c1305af0
- NewGVN: Allow dependent PHI of ops · 9b926e90
  Daniel Berlin authored Sep 30, 2017
  9b926e90