Page MenuHomePhabricator

GBuella (Gabor Buella)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 16 2018, 3:21 AM (74 w, 4 d)

Recent Activity

Feb 27 2019

GBuella added a reviewer for D58619: [NFC][cmake] Build fix in tools/llvm-config/CMakeLists.txt: mgorny.
Feb 27 2019, 2:07 AM · Restricted Project

Feb 25 2019

GBuella created D58619: [NFC][cmake] Build fix in tools/llvm-config/CMakeLists.txt.
Feb 25 2019, 7:04 AM · Restricted Project

Feb 8 2019

GBuella committed rG53980b24b7d0: Extra processing for BitCast + PHI in InstCombine (authored by GBuella).
Extra processing for BitCast + PHI in InstCombine
Feb 8 2019, 5:47 PM
GBuella committed rL353595: Extra processing for BitCast + PHI in InstCombine.
Extra processing for BitCast + PHI in InstCombine
Feb 8 2019, 5:46 PM
GBuella closed D57053: Extra processing for BitCast + PHI in InstCombine.
Feb 8 2019, 5:46 PM · Restricted Project

Feb 7 2019

GBuella closed D57297: [X86] New blsr test cases.
Feb 7 2019, 3:31 PM · Restricted Project

Jan 26 2019

GBuella committed rL352306: [X86] Add some missing blsr patterns.
[X86] Add some missing blsr patterns
Jan 26 2019, 10:16 PM
GBuella closed D57296: [X86] Add some missing blsr patterns.
Jan 26 2019, 10:15 PM
GBuella committed rL352305: [NFC][X86] Add a few more blsr test cases.
[NFC][X86] Add a few more blsr test cases
Jan 26 2019, 10:05 PM
GBuella abandoned D47912: [CMake] Consider LLVM_APPEND_VC_REV when generating SVNVersion.inc.
Jan 26 2019, 9:40 PM
GBuella added a comment to D57296: [X86] Add some missing blsr patterns.

The two corresponding test cases are not yet commited,
https://reviews.llvm.org/D57297

Jan 26 2019, 9:37 PM
GBuella created D57297: [X86] New blsr test cases.
Jan 26 2019, 9:35 PM · Restricted Project
GBuella created D57296: [X86] Add some missing blsr patterns.
Jan 26 2019, 9:35 PM

Jan 16 2019

GBuella committed rL351325: Assertion in isAllocaPromotable due to extra bitcast goes into lifetime marker.
Assertion in isAllocaPromotable due to extra bitcast goes into lifetime marker
Jan 16 2019, 4:10 AM
GBuella closed D55934: Assertion in isAllocaPromotable due to extra bitcast goes into lifetime marker.
Jan 16 2019, 4:10 AM

Oct 25 2018

GBuella committed rL345248: Add -instcombine-code-sinking option.
Add -instcombine-code-sinking option
Oct 25 2018, 1:34 AM
GBuella closed D52709: Add -instcombine-code-sinking option.
Oct 25 2018, 1:34 AM

Oct 22 2018

GBuella added a comment to D52709: Add -instcombine-code-sinking option.

Can you update this review with a summary that describes the problem your are trying to fix by disabling instruction sinking here?

Oct 22 2018, 3:57 AM

Oct 19 2018

GBuella updated the diff for D52709: Add -instcombine-code-sinking option.

Back to having only a CLI option, no constructor argument.

Oct 19 2018, 1:57 AM

Oct 12 2018

GBuella updated the diff for D52709: Add -instcombine-code-sinking option.
Oct 12 2018, 5:41 AM
GBuella updated the diff for D52709: Add -instcombine-code-sinking option.

Adding constructor argument as well -- making it more similar to the ExpensiveCombines option.

Oct 12 2018, 5:18 AM

Oct 1 2018

GBuella added a reviewer for D52709: Add -instcombine-code-sinking option: majnemer.
Oct 1 2018, 1:36 AM
GBuella created D52709: Add -instcombine-code-sinking option.
Oct 1 2018, 1:19 AM

Jul 7 2018

GBuella committed rL336484: NFC - Typo fixes in X86 flags-copy-lowering.mir test.
NFC - Typo fixes in X86 flags-copy-lowering.mir test
Jul 7 2018, 9:14 AM
GBuella closed D48934: NFC - Typo fixes in X86 flags-copy-lowering.mir test.
Jul 7 2018, 9:14 AM

Jul 6 2018

GBuella retitled D48921: NFC - Typo fix in test/CodeGenCXX/runtime-dllstorage.cpp from NFC - type fix in test/CodeGenCXX/runtime-dllstorage.cpp to NFC - Typo fix in test/CodeGenCXX/runtime-dllstorage.cpp.
Jul 6 2018, 2:02 AM
GBuella requested review of D48921: NFC - Typo fix in test/CodeGenCXX/runtime-dllstorage.cpp.

Well, apparently the test fails with the typo fix.
There is no declare dllimport void @_ZSt9terminatev() line that could be matched for CHECK-DYNAMIC-IA-DAG.

Jul 6 2018, 2:00 AM

Jul 5 2018

GBuella abandoned D48487: [X86][AVX512] Lowering _mm512_[max|min]_p[s|d] to native IR.
Jul 5 2018, 2:50 PM
GBuella committed rL336355: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
[X86] Fix some vector cmp builtins - TRUE/FALSE predicates
Jul 5 2018, 7:32 AM
GBuella committed rC336355: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
[X86] Fix some vector cmp builtins - TRUE/FALSE predicates
Jul 5 2018, 7:31 AM
GBuella closed D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
Jul 5 2018, 7:31 AM
GBuella added a comment to D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.

ping @spatel

Jul 5 2018, 6:10 AM
GBuella updated the diff for D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.

As suggested, I added test cases with all predicates (in r336346).

Jul 5 2018, 6:09 AM
GBuella committed rC336346: [X86] NFC - add more test cases for vector cmp intrinsics.
[X86] NFC - add more test cases for vector cmp intrinsics
Jul 5 2018, 6:02 AM
GBuella committed rL336346: [X86] NFC - add more test cases for vector cmp intrinsics.
[X86] NFC - add more test cases for vector cmp intrinsics
Jul 5 2018, 6:02 AM

Jul 4 2018

GBuella committed rL336268: NFC - Various typo fixes in tests.
NFC - Various typo fixes in tests
Jul 4 2018, 6:33 AM
GBuella added a comment to D48934: NFC - Typo fixes in X86 flags-copy-lowering.mir test.

I wasn't 100% sure about these being the intended checks.

Jul 4 2018, 6:20 AM
GBuella created D48934: NFC - Typo fixes in X86 flags-copy-lowering.mir test.
Jul 4 2018, 6:19 AM
GBuella committed rC336264: NFC - Fix type in builtins-ppc-p9vector.c test.
NFC - Fix type in builtins-ppc-p9vector.c test
Jul 4 2018, 4:34 AM
GBuella committed rL336264: NFC - Fix type in builtins-ppc-p9vector.c test.
NFC - Fix type in builtins-ppc-p9vector.c test
Jul 4 2018, 4:34 AM
GBuella committed rL336263: NFC - Fix typo in test/CodeGenObjC/gnustep2-class.m.
NFC - Fix typo in test/CodeGenObjC/gnustep2-class.m
Jul 4 2018, 4:31 AM
GBuella committed rC336263: NFC - Fix typo in test/CodeGenObjC/gnustep2-class.m.
NFC - Fix typo in test/CodeGenObjC/gnustep2-class.m
Jul 4 2018, 4:31 AM
GBuella committed rC336262: NFC - Fix typo in test/Layout/itanium-pack-and-align.cpp.
NFC - Fix typo in test/Layout/itanium-pack-and-align.cpp
Jul 4 2018, 4:26 AM
GBuella committed rL336262: NFC - Fix typo in test/Layout/itanium-pack-and-align.cpp.
NFC - Fix typo in test/Layout/itanium-pack-and-align.cpp
Jul 4 2018, 4:26 AM
GBuella created D48921: NFC - Typo fix in test/CodeGenCXX/runtime-dllstorage.cpp.
Jul 4 2018, 2:05 AM
GBuella committed rC336243: NFC - typo fix in test/CodeGen/avx512f-builtins.c.
NFC - typo fix in test/CodeGen/avx512f-builtins.c
Jul 4 2018, 1:37 AM
GBuella committed rL336243: NFC - typo fix in test/CodeGen/avx512f-builtins.c.
NFC - typo fix in test/CodeGen/avx512f-builtins.c
Jul 4 2018, 1:37 AM

Jun 28 2018

GBuella added inline comments to D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
Jun 28 2018, 11:36 AM
GBuella added inline comments to D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
Jun 28 2018, 7:48 AM
GBuella abandoned D48708: NFC Build fix in RegisterCustomCheckersTest.cpp.
Jun 28 2018, 6:54 AM
GBuella created D48715: [X86] Fix some vector cmp builtins - TRUE/FALSE predicates.
Jun 28 2018, 2:34 AM
GBuella created D48708: NFC Build fix in RegisterCustomCheckersTest.cpp.
Jun 28 2018, 1:16 AM

Jun 22 2018

GBuella created D48487: [X86][AVX512] Lowering _mm512_[max|min]_p[s|d] to native IR.
Jun 22 2018, 8:08 AM
GBuella committed rL335340: [X86] Add notes to a few intrinsics.
[X86] Add notes to a few intrinsics
Jun 22 2018, 5:06 AM
GBuella closed D48280: [X86] Add notes to a few intrinsics.
Jun 22 2018, 5:06 AM
GBuella committed rL335339: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
[X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR
Jun 22 2018, 5:04 AM
GBuella committed rC335339: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
[X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR
Jun 22 2018, 5:04 AM
GBuella closed D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 22 2018, 5:03 AM

Jun 20 2018

GBuella added a comment to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

I was overzealous with the intrinsics, I lower really only the packed comparison now.

Jun 20 2018, 4:54 AM
GBuella updated the diff for D48280: [X86] Add notes to a few intrinsics.
Jun 20 2018, 4:53 AM
GBuella updated the diff for D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 20 2018, 4:52 AM

Jun 19 2018

GBuella updated the diff for D48280: [X86] Add notes to a few intrinsics.

Added __builtin_ia32_cmpsd_mask & __builtin_ia32_cmpss_mask.

Jun 19 2018, 3:25 AM
GBuella updated the diff for D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

Added __builtin_ia32_cmpsd_mask & __builtin_ia32_cmpss_mask.

Jun 19 2018, 3:23 AM

Jun 18 2018

GBuella updated the diff for D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 18 2018, 7:08 AM
GBuella created D48280: [X86] Add notes to a few intrinsics.
Jun 18 2018, 7:04 AM
GBuella added a comment to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

The question still left is, should we remove, auto upgrade the LLVM intrinsics not used anymore, or keep them around for when the signal behaviour is going to matter?

Jun 18 2018, 6:57 AM

Jun 14 2018

GBuella added inline comments to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 14 2018, 12:43 AM

Jun 13 2018

GBuella updated the diff for D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

Ignoring signaling behviour - and rounding mode with it.
Also lowering __builtin_ia32_cmpsd and __builtin_ia32_cmpss.

Jun 13 2018, 11:07 AM
GBuella added inline comments to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 13 2018, 10:11 AM

Jun 12 2018

GBuella added inline comments to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.
Jun 12 2018, 11:18 AM

Jun 11 2018

GBuella added a comment to D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

Ping @efriedma

Jun 11 2018, 8:35 AM
GBuella updated the diff for D45616: [X86] Lower _mm[256|512]_cmp[.]_mask intrinsics to native llvm IR.

I altered the code, to ignore the the signaling behaviour, as suggested.
Also, it handles more vector cmp builtins now.

Jun 11 2018, 8:34 AM
GBuella added inline comments to D46957: [x86] Lower some trunc + shuffle patterns to vpmov[q|d][b|w].
Jun 11 2018, 6:06 AM
GBuella added inline comments to D46957: [x86] Lower some trunc + shuffle patterns to vpmov[q|d][b|w].
Jun 11 2018, 3:44 AM

Jun 9 2018

GBuella committed rL334345: [X86] NFC Use member initialization in X86Subtarget.
[X86] NFC Use member initialization in X86Subtarget
Jun 9 2018, 2:24 AM

Jun 7 2018

GBuella committed rL334247: NFC Fix a comment in ValueTypes.td.
NFC Fix a comment in ValueTypes.td
Jun 7 2018, 4:36 PM
GBuella added a comment to D47912: [CMake] Consider LLVM_APPEND_VC_REV when generating SVNVersion.inc.

I believe LLVM_APPEND_VC_REV controls whether the revision is appended to LLVM version string (e.g. the output of llvm-config --version), whereas the SVNRevision.inc file (which is what's causing the relink after amending) is used for e.g. the clang --version output, so I'm not sure this is the right thing to do. I would also love for an option to disable the SVNRevision stuff entirely though.

Jun 7 2018, 3:38 PM
GBuella created D47912: [CMake] Consider LLVM_APPEND_VC_REV when generating SVNVersion.inc.
Jun 7 2018, 3:21 PM
GBuella committed rC334174: [CodeGen] Improve diagnostics related to target attributes.
[CodeGen] Improve diagnostics related to target attributes
Jun 7 2018, 1:53 AM
GBuella committed rL334174: [CodeGen] Improve diagnostics related to target attributes.
[CodeGen] Improve diagnostics related to target attributes
Jun 7 2018, 1:53 AM
GBuella closed D46541: [CodeGen] Improve diagnostics related to target attributes.
Jun 7 2018, 1:52 AM

Jun 5 2018

GBuella added inline comments to D46541: [CodeGen] Improve diagnostics related to target attributes.
Jun 5 2018, 11:49 AM
GBuella updated the diff for D46541: [CodeGen] Improve diagnostics related to target attributes.
Jun 5 2018, 11:49 AM
GBuella added inline comments to D46541: [CodeGen] Improve diagnostics related to target attributes.
Jun 5 2018, 8:29 AM
GBuella committed rL334016: [X86] NFC Fix typo introduced in r328016 HSI->HDI.
[X86] NFC Fix typo introduced in r328016 HSI->HDI
Jun 5 2018, 5:59 AM
GBuella committed rL334005: [X86] NFC Refactor some code in InstPrinters.
[X86] NFC Refactor some code in InstPrinters
Jun 5 2018, 3:45 AM
GBuella closed D47682: [X86] NFC Refactor some code in InstPrinters.
Jun 5 2018, 3:45 AM

Jun 3 2018

GBuella updated the diff for D47682: [X86] NFC Refactor some code in InstPrinters.
Jun 3 2018, 2:37 AM
GBuella created D47682: [X86] NFC Refactor some code in InstPrinters.
Jun 3 2018, 2:19 AM

Jun 1 2018

GBuella committed rL333732: NFC Avoid a warning in WasmEHPrepare.cpp.
NFC Avoid a warning in WasmEHPrepare.cpp
Jun 1 2018, 12:52 AM

May 30 2018

GBuella committed rL333555: [X86] Lowering FMA intrinsics to native IR (Clang part).
[X86] Lowering FMA intrinsics to native IR (Clang part)
May 30 2018, 8:32 AM
GBuella committed rC333555: [X86] Lowering FMA intrinsics to native IR (Clang part).
[X86] Lowering FMA intrinsics to native IR (Clang part)
May 30 2018, 8:32 AM
GBuella closed D47444: [X86] Lowering FMA intrinsics to native IR (Clang part).
May 30 2018, 8:32 AM
GBuella committed rL333554: [X86] Lowering FMA intrinsics to native IR (LLVM part).
[X86] Lowering FMA intrinsics to native IR (LLVM part)
May 30 2018, 8:29 AM
GBuella closed D47443: [X86] Lowering FMA intrinsics to native IR (LLVM part).
May 30 2018, 8:29 AM

May 24 2018

GBuella committed rC333256: [x86] invpcid intrinsic.
[x86] invpcid intrinsic
May 24 2018, 11:39 PM
GBuella committed rL333256: [x86] invpcid intrinsic.
[x86] invpcid intrinsic
May 24 2018, 11:38 PM
GBuella closed D47142: [x86] invpcid intrinsic.
May 24 2018, 11:38 PM
GBuella committed rL333255: [x86] invpcid LLVM intrinsic.
[x86] invpcid LLVM intrinsic
May 24 2018, 11:36 PM
GBuella closed D47141: [x86] invpcid LLVM intrinsic.
May 24 2018, 11:36 PM