kzhuravl (Konstantin Zhuravlyov)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 16 2016, 12:13 PM (100 w, 2 h)

Recent Activity

Thu, Dec 28

kzhuravl added inline comments to D41614: [AMDGPU][MC][GFX8][GFX9] Added XNACK_MASK support.
Thu, Dec 28, 9:59 AM

Dec 8 2017

kzhuravl committed rL320194: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
AMDGPU/GCN: Bring processors in sync with AMDGPUUsage
Dec 8 2017, 12:53 PM
kzhuravl closed D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage by committing rL320194: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Dec 8 2017, 12:53 PM
kzhuravl committed rL320176: AMDGPU: Report Arg's Value name in metadata if kernel_arg_name metadata is not….
AMDGPU: Report Arg's Value name in metadata if kernel_arg_name metadata is not…
Dec 8 2017, 11:22 AM
kzhuravl closed D40924: AMDGPU: Report Arg's Value name in metadata if kernel_arg_name metadata is not available by committing rL320176: AMDGPU: Report Arg's Value name in metadata if kernel_arg_name metadata is not….
Dec 8 2017, 11:22 AM

Dec 6 2017

kzhuravl created D40924: AMDGPU: Report Arg's Value name in metadata if kernel_arg_name metadata is not available.
Dec 6 2017, 3:17 PM
kzhuravl added a comment to D23403: OpenCL atomics 1.2 2.0 conversion pre-link pass.

Can this be abandoned/closed?

Dec 6 2017, 3:12 PM · Restricted Project

Nov 30 2017

kzhuravl accepted D40653: [AMDGPU] Convert test/tools/llvm-objdump/AMDGPU/source-lines.ll to amdgiz.

LGTM.

Nov 30 2017, 8:40 AM

Nov 29 2017

kzhuravl accepted D40091: [AMDGPU] Add options for waitcnt pass debugging; add instr count in debug output..

LGTM.

Nov 29 2017, 1:27 PM

Nov 28 2017

kzhuravl committed rL319192: AMDGPU: Add num spilled s/vgprs to metadata.
AMDGPU: Add num spilled s/vgprs to metadata
Nov 28 2017, 9:51 AM
kzhuravl closed D40321: AMDGPU: Add num spilled s/vgprs to metadata by committing rL319192: AMDGPU: Add num spilled s/vgprs to metadata.
Nov 28 2017, 9:51 AM
kzhuravl added a comment to D40321: AMDGPU: Add num spilled s/vgprs to metadata.

LGTM provided it meets the needs of the tools? Have they reviewed?

Nov 28 2017, 9:30 AM
kzhuravl added a comment to D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.

Ping

Nov 28 2017, 8:51 AM
kzhuravl accepted D40051: AMDGPU: Rename Bonaire target to be gfx704; remove gfx800 and make Iceland and Tonga both use gfx802; update target feature handling.

LGTM

Nov 28 2017, 8:50 AM
kzhuravl abandoned D40186: AMDGPU: Enable xnack by default for gfx700/kaveri.

xnack starts >=gfx8

Nov 28 2017, 8:49 AM

Nov 27 2017

kzhuravl updated the diff for D40321: AMDGPU: Add num spilled s/vgprs to metadata.

Address review feedback.

Nov 27 2017, 12:38 PM

Nov 21 2017

kzhuravl created D40321: AMDGPU: Add num spilled s/vgprs to metadata.
Nov 21 2017, 1:49 PM

Nov 20 2017

kzhuravl added inline comments to D40051: AMDGPU: Rename Bonaire target to be gfx704; remove gfx800 and make Iceland and Tonga both use gfx802; update target feature handling.
Nov 20 2017, 9:00 AM
kzhuravl added a comment to D40051: AMDGPU: Rename Bonaire target to be gfx704; remove gfx800 and make Iceland and Tonga both use gfx802; update target feature handling.

Eliminate gfx800 and make Iceland and Tonga both use gfx802 as they use the same code.

Nov 20 2017, 8:57 AM

Nov 17 2017

kzhuravl added a comment to D40051: AMDGPU: Rename Bonaire target to be gfx704; remove gfx800 and make Iceland and Tonga both use gfx802; update target feature handling.
  • Add xnack flag to e_flags.
Nov 17 2017, 12:30 PM
kzhuravl added a dependent revision for D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage: D40186: AMDGPU: Enable xnack by default for gfx700/kaveri.
Nov 17 2017, 10:25 AM
kzhuravl added a dependency for D40186: AMDGPU: Enable xnack by default for gfx700/kaveri: D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Nov 17 2017, 10:25 AM
kzhuravl created D40186: AMDGPU: Enable xnack by default for gfx700/kaveri.
Nov 17 2017, 10:24 AM
kzhuravl requested review of D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Nov 17 2017, 10:23 AM
kzhuravl planned changes to D40045: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Nov 17 2017, 10:23 AM
kzhuravl updated the diff for D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.

Change bonaire to gfx704

Nov 17 2017, 10:22 AM

Nov 16 2017

kzhuravl abandoned D40047: AMDGPU/GCN: Remove xnack from 801 and 810.

Leaving them on by default for APUs.

Nov 16 2017, 2:56 PM

Nov 15 2017

kzhuravl abandoned D39879: AMDGPU: Remove -mamdgpu-debugger-abi option.

Agreed to do it later.

Nov 15 2017, 1:11 PM
kzhuravl added inline comments to D40091: [AMDGPU] Add options for waitcnt pass debugging; add instr count in debug output..
Nov 15 2017, 10:47 AM
kzhuravl added a comment to D40091: [AMDGPU] Add options for waitcnt pass debugging; add instr count in debug output..

This will be pretty useful. Thanks!

Nov 15 2017, 10:46 AM

Nov 14 2017

kzhuravl created D40047: AMDGPU/GCN: Remove xnack from 801 and 810.
Nov 14 2017, 12:07 PM
kzhuravl added a comment to D39879: AMDGPU: Remove -mamdgpu-debugger-abi option.

Ping.

Nov 14 2017, 12:07 PM
kzhuravl created D40046: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Nov 14 2017, 12:04 PM
kzhuravl created D40045: AMDGPU/GCN: Bring processors in sync with AMDGPUUsage.
Nov 14 2017, 12:03 PM

Nov 10 2017

kzhuravl committed rL317920: AMDGPU/NFC: Split Processors.td into GCNProcessors.td and R600Processors.td.
AMDGPU/NFC: Split Processors.td into GCNProcessors.td and R600Processors.td
Nov 10 2017, 12:02 PM
kzhuravl closed D39880: AMDGPU/NFC: Split Processors.td into GCNProcessors.td and R600Processors.td by committing rL317920: AMDGPU/NFC: Split Processors.td into GCNProcessors.td and R600Processors.td.
Nov 10 2017, 12:02 PM
kzhuravl committed rL317917: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature.
AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature
Nov 10 2017, 11:28 AM
kzhuravl closed D39878: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature by committing rL317917: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature.
Nov 10 2017, 11:28 AM
kzhuravl accepted D39887: AMDGPU: AMDGPUUsage.rst minor corrections.

LGTM.

Nov 10 2017, 11:15 AM
kzhuravl committed rL317909: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain.
AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain
Nov 10 2017, 11:10 AM
kzhuravl closed D39877: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain by committing rL317909: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain.
Nov 10 2017, 11:10 AM

Nov 9 2017

kzhuravl created D39880: AMDGPU/NFC: Split Processors.td into GCNProcessors.td and R600Processors.td.
Nov 9 2017, 4:32 PM
kzhuravl added a comment to D39828: AMDGPU: Update code object description.

Will this change be backwards compatible?

Nov 9 2017, 4:31 PM
kzhuravl added dependent revisions for D39877: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain: D39879: AMDGPU: Remove -mamdgpu-debugger-abi option, D39878: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature.
Nov 9 2017, 4:16 PM
kzhuravl added a dependency for D39879: AMDGPU: Remove -mamdgpu-debugger-abi option: D39877: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain.
Nov 9 2017, 4:16 PM
kzhuravl added a dependency for D39878: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature: D39877: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain.
Nov 9 2017, 4:16 PM
kzhuravl created D39879: AMDGPU: Remove -mamdgpu-debugger-abi option.
Nov 9 2017, 4:14 PM
kzhuravl created D39878: AMDGPU: Add -mxnack/-mno-xnack options that set +/-xnack feature.
Nov 9 2017, 4:14 PM
kzhuravl created D39877: AMDGPU/NFC: Move getAMDGPUTargetFeatures to AMDGPU toolchain.
Nov 9 2017, 4:12 PM
kzhuravl added a comment to D39828: AMDGPU: Update code object description.

Does this mean the code object format has changed?

Nov 9 2017, 3:44 PM
kzhuravl accepted D39828: AMDGPU: Update code object description.

LGTM.

Nov 9 2017, 3:26 PM
kzhuravl accepted D39828: AMDGPU: Update code object description.

LGTM.

Nov 9 2017, 3:07 PM
kzhuravl added a reviewer for D39857: [AMDGPU] Late parsed / dependent arguments for AMDGPU kernel attributes: aaron.ballman.

Hi Alex, can you rebase on top of trunk (I think you brought in some extra changes from hcc branch) and upload a full diff?

Nov 9 2017, 11:39 AM · Restricted Project

Nov 8 2017

kzhuravl accepted D39807: [AMDGPU] updated PAL metadata record keys.

LGTM.

Nov 8 2017, 11:30 AM

Nov 2 2017

kzhuravl committed rL317280: AMDGPU: Fix warning discovered by r317266 [-Wunused-private-field].
AMDGPU: Fix warning discovered by r317266 [-Wunused-private-field]
Nov 2 2017, 3:35 PM
kzhuravl committed rL317266: AMDGPU: Remove outdated fixme (it was already fixed).
AMDGPU: Remove outdated fixme (it was already fixed)
Nov 2 2017, 1:48 PM

Nov 1 2017

kzhuravl committed rL317114: AMDGPU: Fix set but not used warnings related to AMDGPUAS.
AMDGPU: Fix set but not used warnings related to AMDGPUAS
Nov 1 2017, 12:13 PM
kzhuravl closed D39499: AMDGPU: Fix set but not used warnings related to AMDGPUAS by committing rL317114: AMDGPU: Fix set but not used warnings related to AMDGPUAS.
Nov 1 2017, 12:13 PM
kzhuravl created D39499: AMDGPU: Fix set but not used warnings related to AMDGPUAS.
Nov 1 2017, 11:41 AM

Oct 26 2017

kzhuravl added a reviewer for D39350: AMDGPU: Add CPUCoherentL2 feature: t-tye.

Is it going to be used with amdhsa os? Where can I find the spec for this?

Oct 26 2017, 4:19 PM
kzhuravl added inline comments to D39255: [AMDGPU] Emit metadata for hidden arguments for kernel enqueue.
Oct 26 2017, 2:59 PM
kzhuravl accepted D39255: [AMDGPU] Emit metadata for hidden arguments for kernel enqueue.

Can we do that in a separate patch? This patch is intended for metadata for enqueue_kernel.

Oct 26 2017, 2:58 PM
kzhuravl committed rL316680: AMDGPU: Commit missing fence-barrier test.
AMDGPU: Commit missing fence-barrier test
Oct 26 2017, 10:54 AM

Oct 25 2017

kzhuravl accepted D39306: Fix CodeGen/AMDGPU/fcanonicalize-elimination.ll on FreeBSD 11.0.

LGTM.

Oct 25 2017, 2:40 PM
kzhuravl committed rL316590: AMDGPU: Cleanup memory legalizer load/store tests.
AMDGPU: Cleanup memory legalizer load/store tests
Oct 25 2017, 10:05 AM
kzhuravl committed rL316586: AMDGPU/NFC: Rename memory legalizer tests:.
AMDGPU/NFC: Rename memory legalizer tests:
Oct 25 2017, 9:46 AM

Oct 24 2017

kzhuravl committed rL316483: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.
LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags
Oct 24 2017, 12:05 PM
kzhuravl closed D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags by committing rL316483: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.
Oct 24 2017, 12:05 PM
kzhuravl added a comment to D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.
In D39140#905422, @ruiu wrote:

How did you create Inputs/amdgpu-kernel-2.o? We generally want to avoid checking in binary files, so if it is created by llvm-mc or something, we'd like to check in the source file instead of the binary file.

Oct 24 2017, 10:39 AM
kzhuravl updated the diff for D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.

Address review feedback.

Oct 24 2017, 10:28 AM
kzhuravl committed rL316460: LLD/ELF: Allow targets to set e_flags.
LLD/ELF: Allow targets to set e_flags
Oct 24 2017, 10:02 AM
kzhuravl closed D39139: LLD/ELF: Allow targets to set e_flags by committing rL316460: LLD/ELF: Allow targets to set e_flags.
Oct 24 2017, 10:02 AM
kzhuravl added a comment to D39139: LLD/ELF: Allow targets to set e_flags.
In D39139#904673, @ruiu wrote:

LGTM, but please apply clang-format-diff before committing. Thanks!

Oct 24 2017, 10:01 AM
kzhuravl accepted D39219: AMDGPU: Add max-mix-insts subtarget feature.

LGTM

Oct 24 2017, 7:28 AM

Oct 23 2017

kzhuravl added inline comments to D39219: AMDGPU: Add max-mix-insts subtarget feature.
Oct 23 2017, 5:59 PM
kzhuravl committed rL316389: AMDGPU: Initialize WavefrontSize from TD files.
AMDGPU: Initialize WavefrontSize from TD files
Oct 23 2017, 4:03 PM
kzhuravl closed D39205: AMDGPU: Initialize WavefrontSize from TD files by committing rL316389: AMDGPU: Initialize WavefrontSize from TD files.
Oct 23 2017, 4:03 PM
kzhuravl created D39205: AMDGPU: Initialize WavefrontSize from TD files.
Oct 23 2017, 2:07 PM
kzhuravl added inline comments to D39139: LLD/ELF: Allow targets to set e_flags.
Oct 23 2017, 12:56 PM
kzhuravl committed rL316365: LLD: Fix large integer implicitly truncated to unsigned type gcc warning.
LLD: Fix large integer implicitly truncated to unsigned type gcc warning
Oct 23 2017, 12:32 PM
kzhuravl closed D39186: LLD: Fix large integer implicitly truncated to unsigned type warning.

rL316365.

Oct 23 2017, 12:32 PM
kzhuravl updated the diff for D39186: LLD: Fix large integer implicitly truncated to unsigned type warning.

Address review feedback.

Oct 23 2017, 10:05 AM
kzhuravl created D39186: LLD: Fix large integer implicitly truncated to unsigned type warning.
Oct 23 2017, 9:26 AM
kzhuravl accepted D39036: AMDGPU: Fix default range in non-kernel functions.

LGTM.

Oct 23 2017, 9:24 AM

Oct 20 2017

kzhuravl added a dependency for D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags: D39139: LLD/ELF: Allow targets to set e_flags.
Oct 20 2017, 1:20 PM
kzhuravl added a dependent revision for D39139: LLD/ELF: Allow targets to set e_flags: D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.
Oct 20 2017, 1:20 PM
kzhuravl created D39140: LLD/ELF/AMDGPU: Process AMDGPU-specific e_flags.
Oct 20 2017, 1:20 PM
kzhuravl created D39139: LLD/ELF: Allow targets to set e_flags.
Oct 20 2017, 1:20 PM

Oct 19 2017

kzhuravl committed rL316171: AMDGPU/Docs: Fix unreadable characters.
AMDGPU/Docs: Fix unreadable characters
Oct 19 2017, 10:13 AM

Oct 18 2017

kzhuravl accepted D39073: [AMDGPU] Corrections to memory model description..
Oct 18 2017, 3:05 PM
kzhuravl accepted D39073: [AMDGPU] Corrections to memory model description..

LGTM with typos fixed.

Oct 18 2017, 3:01 PM
kzhuravl committed rL316100: AMDGPU/Docs: Make target naming consistent.
AMDGPU/Docs: Make target naming consistent
Oct 18 2017, 10:59 AM
kzhuravl closed D39019: AMDGPU/Docs: Make target naming consistent by committing rL316100: AMDGPU/Docs: Make target naming consistent.
Oct 18 2017, 10:59 AM
kzhuravl committed rL316097: AMDGPU: Rename MaxFlatWorkgroupSize to MaxFlatWorkGroupSize for consistency.
AMDGPU: Rename MaxFlatWorkgroupSize to MaxFlatWorkGroupSize for consistency
Oct 18 2017, 10:31 AM
kzhuravl closed D38957: AMDGPU: Rename MaxFlatWorkgroupSize to MaxFlatWorkGroupSize for consistency by committing rL316097: AMDGPU: Rename MaxFlatWorkgroupSize to MaxFlatWorkGroupSize for consistency.
Oct 18 2017, 10:31 AM

Oct 17 2017

kzhuravl updated the diff for D38957: AMDGPU: Rename MaxFlatWorkgroupSize to MaxFlatWorkGroupSize for consistency.

Rebase.

Oct 17 2017, 3:53 PM
kzhuravl planned changes to D38959: AMDGPU: Don't use TargetStreamer if it has not been initialized.
Oct 17 2017, 2:49 PM
kzhuravl planned changes to D36802: AMDGPU: Cleanup most of the macros.
Oct 17 2017, 2:49 PM
kzhuravl requested changes to D38634: AMDGPU : Custom lowering constrained fps..
Oct 17 2017, 2:09 PM
kzhuravl created D39019: AMDGPU/Docs: Make target naming consistent.
Oct 17 2017, 1:46 PM