vchuravy (Valentin Churavy)
User

Projects

User does not belong to any projects.

User Details

User Since
Jun 10 2015, 8:17 AM (119 w, 4 d)

Recent Activity

Jul 6 2017

vchuravy added a comment to D28990: Align i128 to 16 bytes.

@echristo I noticed that you reverted r294702 shortly after it landed. What needs to be done to get it working?

Jul 6 2017, 6:21 PM

Feb 13 2017

vchuravy abandoned D27296: Don't assume mingw is providing SSP functions.
Feb 13 2017, 6:04 PM

Jan 4 2017

vchuravy updated the diff for D28221: Add check for AVX512 support before assuming skylake processor is SKX..

address review comments

Jan 4 2017, 2:22 PM
vchuravy updated the diff for D28221: Add check for AVX512 support before assuming skylake processor is SKX..

SKX: Defensively check for AVX512

Jan 4 2017, 2:12 PM
vchuravy added a comment to D28221: Add check for AVX512 support before assuming skylake processor is SKX..

@craig.topper should we still have a feature gate for 0x55? SKX is not on the market yet so we can't verify this yet.

Jan 4 2017, 2:10 AM
vchuravy updated the diff for D28221: Add check for AVX512 support before assuming skylake processor is SKX..

Add model id 0x55 as SKX and consolidate 0x4e & 0x5e into SKL

Jan 4 2017, 2:08 AM

Jan 3 2017

vchuravy added a comment to D28221: Add check for AVX512 support before assuming skylake processor is SKX..

Related issue for clang https://llvm.org/bugs/show_bug.cgi?id=27003 shows that the model id for i5-6200U is also 4e. I also grouped the two Skylake ids together.

Jan 3 2017, 1:20 AM
vchuravy updated the diff for D28221: Add check for AVX512 support before assuming skylake processor is SKX..

groups the model id for skylake and checks for AVX512 to differentiate between SKL and SKX

Jan 3 2017, 1:17 AM

Jan 2 2017

vchuravy updated subscribers of D28221: Add check for AVX512 support before assuming skylake processor is SKX..
Jan 2 2017, 6:34 PM
vchuravy retitled D28221: Add check for AVX512 support before assuming skylake processor is SKX. from to Add check for AVX512 support before assuming skylake processor is SKX..
Jan 2 2017, 6:32 PM
vchuravy abandoned D27653: Make check for atomic alignment target dependent.
Jan 2 2017, 6:25 PM

Dec 11 2016

vchuravy retitled D27653: Make check for atomic alignment target dependent from to Make check for atomic alignment target dependent.
Dec 11 2016, 1:38 PM
vchuravy added a reviewer for D27653: Make check for atomic alignment target dependent: craig.topper.
Dec 11 2016, 1:38 PM

Dec 9 2016

vchuravy updated subscribers of D27296: Don't assume mingw is providing SSP functions.
Dec 9 2016, 8:04 AM
vchuravy added a comment to D27610: CMake is funky on detecting Intel 17 as GCC compatible..

btw, would it work with icl.exe on windows?

As far as I know Intel on Windows is capable off handling gnu-style options, but I don't have a test system.

Dec 9 2016, 8:03 AM
vchuravy added a reviewer for D27610: CMake is funky on detecting Intel 17 as GCC compatible.: chapuni.
Dec 9 2016, 3:59 AM
vchuravy retitled D27610: CMake is funky on detecting Intel 17 as GCC compatible. from to CMake is funky on detecting Intel 17 as GCC compatible..
Dec 9 2016, 3:59 AM

Dec 1 2016

vchuravy added a comment to D27296: Don't assume mingw is providing SSP functions.

I should also note the mingw64 provides "libssp" so the MSVCRT functions are not needed on this target.

Dec 1 2016, 5:58 AM
vchuravy added reviewers for D27296: Don't assume mingw is providing SSP functions: craig.topper, etienne.bergeron.
Dec 1 2016, 2:28 AM
vchuravy added a comment to D27296: Don't assume mingw is providing SSP functions.

The Julia project is currently evaluating updating LLVM from 3.7 to 3.9 and we are working on making sure that it will continue to work on all supported platforms.
We currently support Windows through Mingw and encountered LLVM ERROR: Program used external function '__atomic_store' which could not be resolved! in JIT'ed code.
So far we did not need MSVCRT as a runtime library with Mingw, greatly simplifying deployment.

Dec 1 2016, 2:24 AM
vchuravy retitled D27296: Don't assume mingw is providing SSP functions from to Don't assume mingw is providing SSP functions.
Dec 1 2016, 2:11 AM

Oct 31 2016

vchuravy added a comment to D25865: Fix llvm-shlib cmake build.

Since I am new to LLVM I was wondering what the process of landing this is?

Oct 31 2016, 6:50 PM

Oct 23 2016

vchuravy updated the diff for D25865: Fix llvm-shlib cmake build.

Address review comments

Oct 23 2016, 9:08 PM

Oct 21 2016

vchuravy added reviewers for D25865: Fix llvm-shlib cmake build: axw, beanz.
Oct 21 2016, 1:34 AM
vchuravy retitled D25865: Fix llvm-shlib cmake build from to Fix llvm-shlib cmake build.
Oct 21 2016, 1:33 AM

Sep 27 2016

vchuravy updated subscribers of D24951: Import/adapt the SLEEF vector math-function library as an LLVM runtime.
Sep 27 2016, 6:53 AM

Sep 22 2016

vchuravy updated subscribers of D24839: [MC] Prevent out of order HashDirective lexing in AsmLexer..
Sep 22 2016, 12:52 PM
vchuravy updated subscribers of D24839: [MC] Prevent out of order HashDirective lexing in AsmLexer..
Sep 22 2016, 12:52 PM

Sep 20 2016

vchuravy added a comment to D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

@jpienaar Since I don't have commit access, I think somebody else needs to commit this.

Sep 20 2016, 10:23 AM
vchuravy updated the diff for D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

Fix spelling in comment

Sep 20 2016, 10:20 AM
vchuravy added a comment to D9168: [NVPTX] Check if callsite is defined when computing argument allignment.
You don't own revision D9168: "[NVPTX] Check if callsite is defined when
computing argument allignment". Normally, you should only update
revisions you own. You can "Commandeer" this revision from the web
interface if you want to become the owner.
Sep 20 2016, 9:00 AM
vchuravy updated the diff for D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

addresses review comments.

Sep 20 2016, 8:58 AM
vchuravy added a comment to D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

This is my first contribution to llvm so please let me know if I did something wrong in the process

Sep 20 2016, 8:12 AM
vchuravy updated the diff for D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

Rebases the original changes and adds a test-case

Sep 20 2016, 8:11 AM

Sep 19 2016

vchuravy added a comment to D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

@jpienaar are you planning to work on this again? Or should I give it a go?

Sep 19 2016, 5:11 PM
vchuravy added a comment to D9168: [NVPTX] Check if callsite is defined when computing argument allignment.

I have a potential test-case for this, but the patch doesn't apply cleanly to master so I was unable to test if this solves the problem.

Sep 19 2016, 9:30 AM
vchuravy updated subscribers of D9168: [NVPTX] Check if callsite is defined when computing argument allignment.
Sep 19 2016, 9:28 AM