Integrate llvm 20231012 #15163
Merged
Integrate llvm 20231012 #15163
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
… Move remaining affine test cases to its dialect dir (#67921) (Uday Bondhugula on 2023-10-06 08:32:49 +0530) (0 of 78)
…sform] Add a transform.match.operation_empty op to allow s… (#68319) (Nicolas Vasilache on 2023-10-06 09:25:24 +0200) (0 of 77)
…check-next in nvvm-to-llvm test (nfc) (#68326) (Guray Ozen on 2023-10-06 10:37:23 +0200) (0 of 76)
…sform] Fix handling of transitive include in interpreter. (#67560) (Ingo Müller on 2023-10-06 10:56:57 +0200) (0 of 75)
…unused `from` after llvm/llvm-project@7876899 (Christian Sigg on 2023-10-06 11:16:42 +0200) (0 of 74)
…erization] `MaterializeInDestinationOp`: Support memref destinations (#68074) (Matthias Springer on 2023-10-06 11:57:10 +0200) (0 of 73)
… Fix clang-tidy warnings in Affine Utils (Uday Bondhugula on 2023-10-06 15:27:22 +0530) (0 of 72)
…sform] Allow passing various library files to interpreter. (#67120) (Ingo Müller on 2023-10-06 12:52:49 +0200) (0 of 71)
…sform] Provide a minimal set of utils that allow implementing a simple transform dialect interpreter pass (#68330) (Nicolas Vasilache on 2023-10-06 14:11:05 +0200) (0 of 70)
…-Wunused-function in TransformInterpreterPassBase.cpp (NFC) (Jie Fu on 2023-10-06 21:02:44 +0800) (0 of 69)
…orOps] Don't fold extract chains that include dynamic indices (#68333) (Benjamin Maxwell on 2023-10-06 14:32:37 +0100) (0 of 68)
…enMPIRBuilder] Move copyInput to a passed in lambda function and re-order kernel argument load/stores (#68124) (agozillon on 2023-10-06 16:47:27 +0200) (0 of 67)
…or] Constrain patterns: vector.contract -> vector.outerproduct (Andrzej Warzynski on 2023-10-06 16:07:07 +0000) (0 of 66)
…burger] Fix reduce bug in Fraction class and add tests (#68298) (Abhinav271828 on 2023-10-06 21:47:40 +0530) (0 of 66)
…`lower_unpack` when dynamic dimensions are involved (#68423) (qcolombet on 2023-10-06 22:09:58 +0200) (0 of 65)
…se] introduce MapRef, unify conversion/codegen for reader (#68360) (Aart Bik on 2023-10-06 13:42:01 -0700) (0 of 64)
…se] introduce a pass to stage complex sparse operations in… (#68436) (Peiming Liu on 2023-10-06 14:23:18 -0700) (0 of 63)
…][linalg] Apply direct tosa -> linalg Conv2D lowering (#68304) (Jack Frankland on 2023-10-06 17:10:39 -0700) (0 of 62)
…s] Introduce tblgen-to-irdl tool (#66865) (Kunwar Grover on 2023-10-07 10:39:13 +0530) (0 of 61)
…erization] Follow up for #68074 (#68488) (Matthias Springer on 2023-10-07 10:07:17 -0700) (0 of 60)
…erization] Update empty_tensor_elimination transform op (#68497) (Matthias Springer on 2023-10-08 08:46:43 -0700) (0 of 59)
… conversion to LLVMFunc to account better for properties (#67406) (Mehdi Amini on 2023-10-09 00:41:03 -0700) (0 of 58)
…VE] Restructure sources to match ArmSME dialect (NFC) (#68399) (Benjamin Maxwell on 2023-10-09 10:02:55 +0100) (0 of 57)
…ilder] Remove wrapper function in `createTask`, `createTeams` (#67723) (Shraiysh on 2023-10-09 09:20:31 -0400) (0 of 56)
…or] Restore assert and fix typos (#68581) (Andrzej Warzyński on 2023-10-09 14:22:26 +0100) (0 of 55)
…se] replace specialized buffer setup with util code (#68461) (Aart Bik on 2023-10-09 08:50:59 -0700) (0 of 54)
…] Add verifier for `ArgMax` operator (#68410) (Georgios Pinitas on 2023-10-09 08:55:18 -0700) (0 of 53)
…se] move variable into assert to avoid 'unused' error (#68604) (Aart Bik on 2023-10-09 09:11:26 -0700) (0 of 52)
…IR][Transforms] Fix Mem2Reg removal order to respect dominance (#68877) (Christian Ulmann on 2023-10-12 16:47:06 +0200) (0 of 9)
…se] introduce sparse_tensor.reorder_coo operation (#68827) (Peiming Liu on 2023-10-12 09:42:12 -0700) (0 of 8)
…VE] Avoid UBSAN issue with VectorType::Builder (NFC) (Benjamin Maxwell on 2023-10-12 16:52:33 +0000) (0 of 7)
…se] implement sparse_tensor.reorder_coo (#68916) (Peiming Liu on 2023-10-12 13:22:45 -0700) (0 of 6)
…] Fix circular dependency in (#68934) (Guray Ozen on 2023-10-12 16:24:36 -0700) (0 of 5)
…ir][arith] Canonicalization patterns for `arith.select` (#67809)" (hanhanW on 2023-10-12 17:09:14 -0700) (0 of 5)
stellaraccident
requested review from
MaheshRavishankar,
dcaballe,
hanhanW and
rsuderman
as code owners
stellaraccident
added
benchmarks:cuda
Abbreviated Benchmark Summary@ commit 30b96c5aab5a1adf06774cf0c02dc2bfc83bb06e (no previous benchmark results to compare) Raw Latencies
[Top 3 out of 136 results showed] No improved or regressed compilation metrics 🏖️ For more information: |
nicolasvasilache
approved these changes
Oct 13, 2023
Collaborator
Author
|
Bisected the TF failure to: Will just revert it for now and followup post-integrate. |
Collaborator
Author
|
The latest patch fixes the bazel job, and the others looked good before -- so landing. |
…231012 Re-apply LLVM cherry-picks: f29ae3d4957813d218d589ba1da05e44ee6e422a d94b09d8878afbf6bc8de69010af37067180de67 1151cc387a4ff519f2fc70d37137505a448df829 Add LLVM revert for: e29a253c9ebaded53a823def985364392c4ba4ec
MacDue
reviewed
Oct 17, 2023
Collaborator
Author
|
Oops. Stuff for a little crazy when trying to splice the unreverts from last week in. Sorry :/ |
ramiro050
pushed a commit
to ramiro050/iree
that referenced
this pull request
Dec 19, 2023
Carrying local reverts for: * iree-org#15083 (fixes still being made) * llvm/llvm-project#67816 while we sync to a corresponding TF version * llvm/llvm-project#67809, which has already been reverted at HEAD and will drop on its own Landing with local revert: iree-org#15207 --------- Co-authored-by: Kunwar Grover <[email protected]> Co-authored-by: Julian Walker <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Carrying local reverts for:
shiftattribute ofTOSA_MulOpwith the spec llvm/llvm-project#67816 while we sync to a corresponding TF versionarith.selectllvm/llvm-project#67809, which has already been reverted at HEAD and will drop on its ownLanding with local revert: #15207