Today

Prazek added inline comments to D29624: [DSE] Basic MemorySSA-backed global DSE.
Sun, Feb 26, 2:06 AM
Prazek added inline comments to D30062: Estimate speedup due to inlining and use that to adjust threshold..
Sun, Feb 26, 1:29 AM

Yesterday

davide committed rL296294: [LoopDeletion] Modernize and simplify a bit. NFCI..
[LoopDeletion] Modernize and simplify a bit. NFCI.
Sat, Feb 25, 11:20 PM
ctopper committed rL296293: [X86] Fix execution domain for cmpss/sd instructions..
[X86] Fix execution domain for cmpss/sd instructions.
Sat, Feb 25, 10:58 PM
ctopper committed rL296290: [AVX-512] Fix the execution domain for AVX-512 integer broadcasts..
[AVX-512] Fix the execution domain for AVX-512 integer broadcasts.
Sat, Feb 25, 10:57 PM
ctopper committed rL296291: [AVX-512] Fix execution domain for vmovhpd/lpd/hps/lps..
[AVX-512] Fix execution domain for vmovhpd/lpd/hps/lps.
Sat, Feb 25, 10:57 PM
ctopper committed rL296292: [AVX-512] Fix execution domain for scalar commutable min/max instructions..
[AVX-512] Fix execution domain for scalar commutable min/max instructions.
Sat, Feb 25, 10:57 PM
ctopper committed rL296288: [AVX-512] Fix execution domain for VPMADD52 instructions..
[AVX-512] Fix execution domain for VPMADD52 instructions.
Sat, Feb 25, 10:57 PM
ctopper committed rL296287: [AVX-512] Use update_llc_test_checks.py to regenerate a test..
[AVX-512] Use update_llc_test_checks.py to regenerate a test.
Sat, Feb 25, 10:57 PM
ctopper committed rL296289: [AVX-512] Disable the redundant patterns in the VPBROADCASTBr_Alt and….
[AVX-512] Disable the redundant patterns in the VPBROADCASTBr_Alt and…
Sat, Feb 25, 10:57 PM
ctopper committed rL296284: [X86] Fix the execution domain for scalar SQRT intrinsic instruction..
[X86] Fix the execution domain for scalar SQRT intrinsic instruction.
Sat, Feb 25, 10:57 PM
ctopper committed rL296286: [AVX-512] Fix the execution domain for VSCALEF instructions..
[AVX-512] Fix the execution domain for VSCALEF instructions.
Sat, Feb 25, 10:57 PM
ctopper committed rL296285: [AVX-512] Fix execution domain of scalar VRANGE/REDUCE/GETMANT with sae..
[AVX-512] Fix execution domain of scalar VRANGE/REDUCE/GETMANT with sae.
Sat, Feb 25, 10:57 PM
ctopper committed rL296283: [X86] Add an additional CHECK prefix to a test. Some of the cases used it, but….
[X86] Add an additional CHECK prefix to a test. Some of the cases used it, but…
Sat, Feb 25, 10:57 PM
akyrtzi closed D30304: [indexer] Expose the logic for indexing local symbols through c-index-test and add tests for it.
Sat, Feb 25, 9:57 PM
akyrtzi accepted D30304: [indexer] Expose the logic for indexing local symbols through c-index-test and add tests for it.

Committed in r296282 with a couple of changes:

  • Made it so that parameter definitions have a 'child' relation with their function/method (instead of 'containedBy'). This allows looking up parameters of a function by looking up their children.
  • Given that parameters show up in parent-child hierarchies, I think it is a bit better to make them to a separate 'SymbolKind' instead of a sub-kind. This preserves the property that 'variable' symbol kind does not ever enter a parent-child hierarchy.
Sat, Feb 25, 9:56 PM
akirtzidis committed rL296282: [index] Add 'Parameter' symbol kind and 'Local' symbol property to distinguish….
[index] Add 'Parameter' symbol kind and 'Local' symbol property to distinguish…
Sat, Feb 25, 9:50 PM
bmharper added a comment to D21279: Fix some issues in clang-format's AlignConsecutive modes.

Hi @djasper,
This is the first patch I've contributed here, so I'm not familiar with the whole process. I assume this code is ready to land? When exactly does it get merged into master, and is there something else that I still need to do to make that happen?

Sat, Feb 25, 8:12 PM · Restricted Project
trentxintong added a comment to D30381: [SCCP] Remove manual folding of terminator instructions..

@davide Thanks for taking a quick look !

Sat, Feb 25, 7:25 PM
trentxintong updated the diff for D30322: [SCCP] Resolve indirect branch target when possible..

Rebase.

Sat, Feb 25, 7:16 PM
trentxintong committed rL296281: [SCCP] Remove manual folding of terminator instructions..
[SCCP] Remove manual folding of terminator instructions.
Sat, Feb 25, 6:23 PM
trentxintong closed D30381: [SCCP] Remove manual folding of terminator instructions. by committing rL296281: [SCCP] Remove manual folding of terminator instructions..
Sat, Feb 25, 6:23 PM
davide added a comment to D30350: [LSR] Add a cap for reassociation of AllFixupsOutsideLoop type LSRUse to protect compile time.

Also, I think the test from the PR (after some polishing, maybe) is good to add. If somebody removes these lines at least all (or a subset of) the bots will timeout and we'll notice the regression.

Sat, Feb 25, 6:05 PM
davide added inline comments to D30350: [LSR] Add a cap for reassociation of AllFixupsOutsideLoop type LSRUse to protect compile time.
Sat, Feb 25, 6:04 PM
davide accepted D30381: [SCCP] Remove manual folding of terminator instructions..

I actually forgot to recommit this one :(
The issue with the original patch was that I folded the call inside the assert so ConstantFoldTerminator() was compiled away in !DEBUG mode.
Oh, well. Your reasoning is correct (actually, was my original reasoning when I committed the patch).

Sat, Feb 25, 6:02 PM
dlj committed rL296280: [X86] Clean up test/CodeGen/X86/2006-03-02-InstrSchedBug.ll.
[X86] Clean up test/CodeGen/X86/2006-03-02-InstrSchedBug.ll
Sat, Feb 25, 5:44 PM
dlj closed D30320: [X86] Clean up test/CodeGen/X86/2006-03-02-InstrSchedBug.ll by committing rL296280: [X86] Clean up test/CodeGen/X86/2006-03-02-InstrSchedBug.ll.
Sat, Feb 25, 5:44 PM
dlj accepted D30320: [X86] Clean up test/CodeGen/X86/2006-03-02-InstrSchedBug.ll.

Looks good. I will push shortly.

Sat, Feb 25, 5:44 PM
dlj abandoned D29799: First cut at moving driver tools into their own files..
Sat, Feb 25, 5:42 PM
niravd committed rL296279: Revert "In visitSTORE, always use FindBetterChain, rather than only when UseAA….
Revert "In visitSTORE, always use FindBetterChain, rather than only when UseAA…
Sat, Feb 25, 5:40 PM
ruiu committed rL296278: Add a section about how to use ld.lld..
Add a section about how to use ld.lld.
Sat, Feb 25, 4:32 PM
echristo committed rL296277: vec perm can go down either pipeline on P8..
vec perm can go down either pipeline on P8.
Sat, Feb 25, 4:23 PM
davidxl added a comment to D30309: CodeGen: BlockPlacement: Precompute layout for chains of triangles..

if BP is not correct, it is better to improve static branch prediction. We explicitly added a threshold for the cost based analysis result to kick in just to be conservative when the branch probability is not biased enough. Even for the long chain case, tail dup is enabled for 50/50 case, but the real profile is 40/60, taildup will hurt performance. I don't see the reason to by pass the branch prob + cost analysis by just looking at the shape.

Sat, Feb 25, 4:08 PM
rsmith committed rL296276: Update cxx_dr_status page..
Update cxx_dr_status page.
Sat, Feb 25, 4:06 PM
rsmith committed rL296275: C++ DR1611, 1658, 2180: implement "potentially constructed subobject" rules for….
C++ DR1611, 1658, 2180: implement "potentially constructed subobject" rules for…
Sat, Feb 25, 4:05 PM
sanjoy committed rL296274: Fix signed-unsigned comparison warning.
Fix signed-unsigned comparison warning
Sat, Feb 25, 2:37 PM
silvas added a comment to D26872: Outliner: Add MIR-level outlining pass.

btw, down the road you may want to have this pass really know in detail the encoded length of each instruction on x86. There are quite a few *single instructions* that would be beneficial from a code size perspective to outline (if the outlined function is set to have alignment of 1). A quick analysis of an LLD binary (which contains all of LLVM linked in for LTO) shows there is over 5% code size savings just from outlining single instructions (since many x86 instructions encode to be larger than a CALL instruction which is 5 bytes). About half of the benefit (so about 2-3% of the total on this test case) comes from instructions that reference the stack via %rsp (mostly zeroing out stack slots), which could still be outlined if the offset was rewritten.

Sat, Feb 25, 2:37 PM
efriedma added a comment to D29845: [SelectionDAG] Remove redundant stores more aggressively..

Are there other cases that you've noticed?

Sat, Feb 25, 2:33 PM
axeldavy created D30382: AMDGPU/SI: Remove double lookups in SI scheduler.
Sat, Feb 25, 2:19 PM · Restricted Project
mehdi_amini accepted D30380: Teach lit to expand glob expressions.

I believe you :)

Sat, Feb 25, 2:16 PM