zvi (Zvi Rackover)
User

Projects

User does not belong to any projects.

User Details

User Since
Jun 8 2016, 12:50 PM (76 w, 38 m)

Recent Activity

Today

zvi added a comment to D40330: Separate ExecutionDepsFix into 3 parts - enable breaking false dependencies for all reg classes..

LGTM with a minor comment. Thanks!

Wed, Nov 22, 9:26 AM
zvi added a comment to D40334: [X86] Break false dependencies for POPCNT, LZCNT, TZCNT.

I think we will need a subtarget feature indicating whether the false-dependency on these instructions happens. For starters, does this happen in AMD processors?

Wed, Nov 22, 9:24 AM
zvi added inline comments to D40290: [X86] Fix a bug in handling GRXX subclasses in Domain Reassignment pass.
Wed, Nov 22, 8:02 AM

Mon, Nov 20

zvi added a comment to D40231: InstructionSimplify: 'extractelement' with an undef index is undef.

One option is to run llc with -O0 which will disable passes which use InstructionSimplify, but this will require changing the CHECK's to expect different generated code. If the purpose of the test is to not crash in the Verifier, are the CHECK's needed?

Mon, Nov 20, 11:00 PM
zvi added inline comments to D39421: [InstCombine] Extracting common and-mask for shift operands of Or instruction.
Mon, Nov 20, 11:00 AM

Sun, Nov 19

zvi added a comment to D40231: InstructionSimplify: 'extractelement' with an undef index is undef.

With this patch i am seeing failures in test/CodeGen/AMDGPU/indirect-addressing-si.ll . This test include extractelement instructions with undef indices, which is what this patch targets. I would appreciate your help with advice on how we can modify these tests so that they can be used safely. Please bear in mind that i have no knowledge of the AMDGPU backend.

Sun, Nov 19, 3:07 PM
zvi created D40231: InstructionSimplify: 'extractelement' with an undef index is undef.
Sun, Nov 19, 3:01 PM
zvi added inline comments to D39952: [X86]: Adding full coverage of MC encoding for all X86 ISA Sets.<NFC>.
Sun, Nov 19, 12:41 PM
zvi updated the diff for D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Rebase on ToT. NFC in this revision.

Sun, Nov 19, 12:24 PM
zvi updated the diff for D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Address the last of Craig's comments:

  • Thanks, @lsaba, for porting the pass to the new PassManager.
  • Removed shrinkage of vector types until we sort out if it is generally allowed to shrink element types of vector operations.
  • Some minor fixes to comments.
Sun, Nov 19, 12:15 PM
zvi added inline comments to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.
Sun, Nov 19, 12:51 AM

Thu, Nov 16

zvi updated subscribers of D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.
Thu, Nov 16, 6:31 AM
zvi updated the diff for D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Address some of Craig's recent comments.

Thu, Nov 16, 4:47 AM
zvi added inline comments to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.
Thu, Nov 16, 4:45 AM

Sun, Nov 12

zvi added inline comments to D39840: [MC][X86] Code padding for performance stability - Branch instructions and targets alignment.
Sun, Nov 12, 11:20 PM

Thu, Nov 9

zvi commandeered D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Commandeering this patch while Amjad is away on a few weeks of vacation.
Also sending a friendly ping to the reviewers. AFAIK, all comments were addressed as of the latest revision of this patch. Please let me know if i missed anything. Thanks.

Thu, Nov 9, 12:44 AM

Tue, Nov 7

zvi added a comment to D38417: [test-suite] Adding HACCKernels app.

Since the final commit of this patch, rL317483, the AVX2 buildbot is broken: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/1402

Tue, Nov 7, 10:17 PM

Mon, Nov 6

zvi committed rL317463: X86 ISel: Basic support for variable-index vector permutations.
X86 ISel: Basic support for variable-index vector permutations
Mon, Nov 6, 12:26 AM
zvi closed D39126: X86 ISel: Basic support for variable-index vector permutations by committing rL317463: X86 ISel: Basic support for variable-index vector permutations.
Mon, Nov 6, 12:26 AM
zvi updated the diff for D39126: X86 ISel: Basic support for variable-index vector permutations.

Last minute changes in the spirit of Simon's suggestion

Mon, Nov 6, 12:25 AM

Thu, Nov 2

zvi updated the diff for D39126: X86 ISel: Basic support for variable-index vector permutations.

Follow Simon's suggestion to pull the oth index scan into the loop.
Fix typo + style defects.

Thu, Nov 2, 3:37 PM
zvi added a comment to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Thanks, Amjad! This patch LGTM, but i think it would be best to wait for an LGTM from one of the assigned reviewers.

Thu, Nov 2, 8:32 AM
zvi added a comment to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Some more minor comments

Thu, Nov 2, 7:07 AM

Wed, Nov 1

zvi added a comment to D39476: [X86][SSE] Truncate with PACKSS any input with sufficient sign-bits.

LGTM

Wed, Nov 1, 2:31 AM

Mon, Oct 30

zvi updated the diff for D39126: X86 ISel: Basic support for variable-index vector permutations.

Add support for floating-point types.

Mon, Oct 30, 2:20 PM
zvi committed rL316946: X86 Tests: Update the variable-index permute tests with FP types. NFC..
X86 Tests: Update the variable-index permute tests with FP types. NFC.
Mon, Oct 30, 12:29 PM
zvi added a comment to D39126: X86 ISel: Basic support for variable-index vector permutations.

ping

Mon, Oct 30, 4:37 AM
zvi added a comment to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

I know you decided to move this to a new pass, but here's a couple of more comments that will be relevant.

Mon, Oct 30, 4:32 AM
zvi accepted D38684: [X86][AVX512] lowering broadcastm intrinsic - llvm part.

LGTM

Mon, Oct 30, 4:08 AM
zvi accepted D39411: [X86] Make sure we don't create locked inc/dec instructions when the carry flag is being used..

LGTM

Mon, Oct 30, 4:05 AM
zvi added a comment to D39402: [X86] Prevent fast isel from folding loads into the instructions listed in hasPartialRegUpdate..

LGTM, with a minor concern that with this change FastISel is a bit slower because the bail-out condition was moved to a later point?

Mon, Oct 30, 3:14 AM

Sun, Oct 29

zvi added a comment to D37775: Add a verifier test to check the access on both sides of COPY are the same.

Hi All,

Sun, Oct 29, 2:09 AM

Thu, Oct 26

zvi added a comment to D38313: [InstCombine] Introducing Aggressive Instruction Combine pass.

Amjad, some questions about where do we want this work to evolve to.

Thu, Oct 26, 6:45 AM

Tue, Oct 24

zvi added inline comments to D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Tue, Oct 24, 3:25 PM
zvi committed rL316434: X86CallFrameOptimization: Update comments and variable names. NFCI..
X86CallFrameOptimization: Update comments and variable names. NFCI.
Tue, Oct 24, 6:25 AM
zvi committed rL316431: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms
Tue, Oct 24, 5:13 AM
zvi closed D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms by committing rL316431: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
Tue, Oct 24, 5:13 AM
zvi added inline comments to D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
Tue, Oct 24, 4:56 AM
zvi committed rL316416: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
X86: Fix X86CallFrameOptimization to search for the COPY StackPointer
Tue, Oct 24, 12:39 AM
zvi added inline comments to D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Tue, Oct 24, 12:39 AM
zvi closed D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer by committing rL316416: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Tue, Oct 24, 12:38 AM

Mon, Oct 23

zvi committed rL316412: X86: Register the X86CallFrameOptimization pass.
X86: Register the X86CallFrameOptimization pass
Mon, Oct 23, 10:48 PM
zvi closed D38729: X86: Register the X86CallFrameOptimization pass by committing rL316412: X86: Register the X86CallFrameOptimization pass.
Mon, Oct 23, 10:48 PM

Oct 20 2017

zvi created D39126: X86 ISel: Basic support for variable-index vector permutations.
Oct 20 2017, 8:39 AM
zvi committed rL316216: X86 Tests: Add tests for vector permutes with variable indices. NFC..
X86 Tests: Add tests for vector permutes with variable indices. NFC.
Oct 20 2017, 8:32 AM

Oct 19 2017

zvi added a comment to D39077: [X86] Teach the assembly parser to warn on duplicate registers in gather instructions..

LGTM

Oct 19 2017, 4:35 AM
zvi added a reviewer for D39077: [X86] Teach the assembly parser to warn on duplicate registers in gather instructions.: coby.
Oct 19 2017, 4:28 AM

Oct 18 2017

zvi added a comment to D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.

ping

Oct 18 2017, 6:30 AM
zvi added a comment to D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.

ping

Oct 18 2017, 6:30 AM
zvi added a comment to D38729: X86: Register the X86CallFrameOptimization pass.

ping

Oct 18 2017, 6:29 AM

Oct 17 2017

zvi added a comment to D36597: DAG: Fix creating select with wrong condition type.

The changes in WidenVSELECTAndMask and the X86 test LGTM.

Oct 17 2017, 8:39 AM

Oct 16 2017

zvi accepted D37251: [X86] Add a pass to convert instruction chains between domains.

LGTM

Oct 16 2017, 1:16 PM

Oct 10 2017

zvi added a dependency for D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms: D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Oct 10 2017, 8:18 AM
zvi added a dependent revision for D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer: D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
Oct 10 2017, 8:18 AM
zvi updated the summary of D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
Oct 10 2017, 8:18 AM
zvi created D38738: X86CallFrameOptimization: Recognize 'store 0/-1 using and/or' idioms.
Oct 10 2017, 8:17 AM
zvi added a dependency for D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer: D38729: X86: Register the X86CallFrameOptimization pass.
Oct 10 2017, 7:21 AM
zvi added a dependent revision for D38729: X86: Register the X86CallFrameOptimization pass: D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Oct 10 2017, 7:21 AM
zvi added inline comments to D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Oct 10 2017, 7:20 AM
zvi created D38730: X86: Fix X86CallFrameOptimization to search for the COPY StackPointer.
Oct 10 2017, 6:57 AM
zvi created D38729: X86: Register the X86CallFrameOptimization pass.
Oct 10 2017, 6:46 AM

Oct 9 2017

zvi committed rL315237: [X86] Unsigned saturation subtraction canonicalization [the backend part].
[X86] Unsigned saturation subtraction canonicalization [the backend part]
Oct 9 2017, 1:03 PM
zvi closed D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part] by committing rL315237: [X86] Unsigned saturation subtraction canonicalization [the backend part].
Oct 9 2017, 1:03 PM
zvi accepted D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part].

LGTM. Thanks!

Oct 9 2017, 12:13 AM

Oct 1 2017

zvi accepted D37687: [X86][LLVM]Expanding Supports lowerInterleaved{store|load}() in X86InterleavedAccess (VF64 stride 3) .

LGTM, and yes please add the assert

Oct 1 2017, 12:54 AM

Sep 28 2017

zvi added inline comments to D37687: [X86][LLVM]Expanding Supports lowerInterleaved{store|load}() in X86InterleavedAccess (VF64 stride 3) .
Sep 28 2017, 8:54 AM
zvi added inline comments to D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part].
Sep 28 2017, 7:09 AM
zvi accepted D38305: [X86] Use BWI instructions to improve lowering of v32i8 MULHU/S.

LGTM

Sep 28 2017, 5:56 AM
zvi added a comment to D38276: [X86] Remove dead code from X86ISelDAGToDAG.cpp multiply handling.

Rough outline of mul handling as I understand it.

Sep 28 2017, 5:50 AM
zvi accepted D38276: [X86] Remove dead code from X86ISelDAGToDAG.cpp multiply handling.

LGTM

Sep 28 2017, 5:50 AM

Sep 27 2017

zvi added inline comments to D38318: [X86][SSE] Match PSHUFLW/PSHUFHW + PSHUFD vXi16 shuffle patterns (PR34686).
Sep 27 2017, 9:47 AM
zvi committed rL314305: X86 Tests: Unsigned saturation subtraction tests. NFC..
X86 Tests: Unsigned saturation subtraction tests. NFC.
Sep 27 2017, 7:39 AM
zvi closed D37510: Unsigned saturation subtraction canonicalization by committing rL314305: X86 Tests: Unsigned saturation subtraction tests. NFC..
Sep 27 2017, 7:39 AM
zvi accepted D38222: [X86][SKX][KNL] Updated regression tests to use -mattr instead of -mcpu flag.NFC..

LGTM

Sep 27 2017, 6:00 AM
zvi added inline comments to D38312: adding pattern for broadcastm.
Sep 27 2017, 5:47 AM
zvi added inline comments to D38312: adding pattern for broadcastm.
Sep 27 2017, 5:46 AM
zvi accepted D37510: Unsigned saturation subtraction canonicalization.

LGTM

Sep 27 2017, 5:42 AM
zvi added a comment to D38312: adding pattern for broadcastm.

The title should have an X86 label, and can you please add a cover message in the summary?

Sep 27 2017, 5:38 AM
zvi added a comment to D38219: [GlobalISel][X86] Support G_LSHR/G_ASHR/G_SHL..

LGTM with minor comments

Sep 27 2017, 4:08 AM
zvi added inline comments to D37510: Unsigned saturation subtraction canonicalization.
Sep 27 2017, 3:22 AM
zvi added a comment to D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part].

Please add the avx512vl feature to the AVX512 test (in the parent patch) and rebase this patch. It will drastically improve some of the cases :).

Sep 27 2017, 2:43 AM

Sep 25 2017

zvi accepted D37117: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF{8|16|32} stride 3)...

LGTM

Sep 25 2017, 5:26 AM
zvi accepted D36058: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF8 stride 4)..

LGTM, thanks!

Sep 25 2017, 5:03 AM
zvi added inline comments to D38222: [X86][SKX][KNL] Updated regression tests to use -mattr instead of -mcpu flag.NFC..
Sep 25 2017, 5:03 AM
zvi added inline comments to D37117: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF{8|16|32} stride 3)...
Sep 25 2017, 5:03 AM
zvi added a comment to D37117: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF{8|16|32} stride 3)...

Please remember to create the diff with -U

Sep 25 2017, 5:03 AM

Sep 13 2017

zvi added a comment to D37796: [X86][Skylake][KNL] Updating code gen regression test to use the KNL and SKYLAKE prefixes. NFC,.

Can you please move the NFC to the Title:
[X86][Skylake][KNL] Updating code gen regression test to use the KNL and SKYLAKE prefixes. NFC.

Sep 13 2017, 1:21 AM

Sep 12 2017

zvi accepted D34816: [GlobalISel][X86] support G_FPEXT operation..

LGTM

Sep 12 2017, 11:43 PM
zvi accepted D37751: [X86 CodeGen] Optimization of ZeroExtendLoad for v2i8 vector.

LGTM, and as you mentioned in the cover message, there is more room for improvement.

Sep 12 2017, 11:04 PM
zvi added a reviewer for D34596: [X86]: Adding a new priority function 'guided-src' for Scheduler DAG instruction scheduling.: myatsina.
Sep 12 2017, 5:13 AM
zvi added inline comments to D36058: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF8 stride 4)..
Sep 12 2017, 5:06 AM
zvi added inline comments to D37687: [X86][LLVM]Expanding Supports lowerInterleaved{store|load}() in X86InterleavedAccess (VF64 stride 3) .
Sep 12 2017, 2:13 AM

Sep 11 2017

zvi updated subscribers of D37702: [LV] Clamp the VF to the trip count.
Sep 11 2017, 1:09 PM
zvi added a comment to D36058: [X86][LLVM]Expanding Supports lowerInterleavedStore() in X86InterleavedAccess (VF8 stride 4)..

If this patch is still in review please rebase it.

Sep 11 2017, 12:22 PM
zvi added inline comments to D37510: Unsigned saturation subtraction canonicalization.
Sep 11 2017, 12:10 PM
zvi committed rL312921: X86 Tests: More AVX512 conversions tests. NFC.
X86 Tests: More AVX512 conversions tests. NFC
Sep 11 2017, 8:56 AM
zvi added a comment to D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part].

Please add some details about the combine in the Summary. It's better to put a description in addition to the links to make it easier to read the commit log.
Maybe something like:
Combine:
umax(a,b) - b -> subus(a,b)
a - umin(a,b) -> subus(a,b)

Sep 11 2017, 7:20 AM
zvi added a comment to D37510: Unsigned saturation subtraction canonicalization.

Looks ready apart from the missing AVX512 testing Dave suggested.

Sep 11 2017, 6:48 AM
zvi added reviewers for D37534: [X86] Unsigned saturation subtraction canonicalization [the backend part]: zvi, spatel, DavidKreitzer.
Sep 11 2017, 6:43 AM

Sep 7 2017

zvi added a comment to D37451: [x86] enable storeOfVectorConstantIsCheap() target hook.

Intuitively, it seems to me that choosing a minimum threshold, as suggested in note 2, is a better option.
Another concern for store-merging in general: are we more susceptible to losing store-to-load forwarding? I know that Intel pre-Nehalem processors had some limitations that were later improved. Sorry i can't recall the full details from the top of my head. Will look later at the Optimization Manual for the info.

Sep 7 2017, 8:59 AM