Page MenuHomePhabricator

airlied (Dave Airlie)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 9 2016, 11:45 PM (223 w, 3 d)

Recent Activity

Nov 26 2020

airlied updated the diff for D92004: [OpenCL] add CL 3.0 optional feature support to opencl-c.h.

I've moved a bunch of longer checks into just feature checks and defined the features at the top of the file.

Nov 26 2020, 7:16 PM · Restricted Project

Nov 24 2020

airlied added a comment to D92004: [OpenCL] add CL 3.0 optional feature support to opencl-c.h.

Btw how about making some checks simpler. We could always define feature macros __opencl_c_atomic_scope_device, __opencl_c_generic_address_space for OpenCL 2.0 or C++ for OpenCL . Then everywhere we would only need to check feature macros instead of language versions and feature macros together.

#if !(defined()) && (defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ == CL_VERSION_2_0))
#define __opencl_c_atomic_scope_device      1
#define __opencl_c_generic_address_space   1
...
#endif
Nov 24 2020, 12:30 PM · Restricted Project

Nov 23 2020

airlied added a comment to D92004: [OpenCL] add CL 3.0 optional feature support to opencl-c.h.

this file has never been clang-format clean, happy to make changes if reviewer decides they are needed.

Nov 23 2020, 10:41 PM · Restricted Project
airlied requested review of D92004: [OpenCL] add CL 3.0 optional feature support to opencl-c.h.
Nov 23 2020, 9:44 PM · Restricted Project

Oct 5 2020

airlied added a comment to D88467: fix out-of-tree clang build due to sysexits change.

I don't have commit access, can someone please push this.

Oct 5 2020, 4:47 PM · Restricted Project, Restricted Project

Sep 28 2020

airlied requested review of D88467: fix out-of-tree clang build due to sysexits change.
Sep 28 2020, 10:02 PM · Restricted Project, Restricted Project

Aug 29 2018

airlied updated the diff for D48582: Reverse subregister saved loops in register usage info collector..

Rebased onto latest master. I require someone else to push this as I don't have commit rights.

Aug 29 2018, 3:25 PM

Jul 9 2018

airlied abandoned D48578: use a MapVector for pass manage LastUser.

Thanks for the review, Chandler,

Jul 9 2018, 4:27 PM

Jun 25 2018

airlied created D48582: Reverse subregister saved loops in register usage info collector..
Jun 25 2018, 8:10 PM
airlied created D48578: use a MapVector for pass manage LastUser.
Jun 25 2018, 7:21 PM

May 1 2018

airlied created D46346: AMDGPU: rename OpenCL lowering pass to be R600 specific..
May 1 2018, 8:15 PM · Restricted Project
airlied abandoned D32413: AMDGPU: try and convert fmad to add + omod.
May 1 2018, 8:13 PM

Jun 25 2017

airlied added a comment to D34549: fix regression on VI with SDWA gfx9 support.

At the moment this is more than an assert fix, we have broken VI in favour of GFX9 which nobody has. Please get a workaround, revert or fix into llvm master asap.

Jun 25 2017, 4:46 PM
airlied accepted D34495: AMDGPU: Partially fix implicit.buffer.ptr intrinsic handling.
Jun 25 2017, 4:45 PM
airlied added a comment to D34495: AMDGPU: Partially fix implicit.buffer.ptr intrinsic handling.

Looks good to me, I regression tested a few things with it in geom/tess shader land and they seemed fine.

Jun 25 2017, 4:45 PM

Jun 22 2017

airlied added a comment to D34549: fix regression on VI with SDWA gfx9 support.

Consider this a bug report for Sam, I'm not even sure where to start writing a test.

Jun 22 2017, 9:02 PM
airlied added a reviewer for D34549: fix regression on VI with SDWA gfx9 support: SamWot.
Jun 22 2017, 8:44 PM
airlied created D34549: fix regression on VI with SDWA gfx9 support.
Jun 22 2017, 8:40 PM

May 16 2017

airlied added a comment to D33114: [AMDGPU] Fixes for the new waitcnt insertion pass. Add test..

radv regression looks related do ds_bpermute interactions if I had to guess.

May 16 2017, 5:36 PM

May 8 2017

airlied added inline comments to D32960: AMDGPU/SI: Add llvm.amdgcn.s.buffer.load intrinsic.
May 8 2017, 6:18 PM
airlied updated the diff for D32960: AMDGPU/SI: Add llvm.amdgcn.s.buffer.load intrinsic.

This adds support for the x2/x4 paths. Probably could be done better, and I've tested this with a few vulkan games.

May 8 2017, 5:13 PM
airlied created D32960: AMDGPU/SI: Add llvm.amdgcn.s.buffer.load intrinsic.
May 8 2017, 12:19 AM

Apr 23 2017

airlied added a reviewer for D32413: AMDGPU: try and convert fmad to add + omod: arsenm.
Apr 23 2017, 9:14 PM
airlied created D32413: AMDGPU: try and convert fmad to add + omod.
Apr 23 2017, 9:14 PM

Jan 23 2017

airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

fixed a regression in the when to emit spill setup path, should have just added isMesaGfx option not removed the other one.

Jan 23 2017, 9:48 PM
airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

fix pedantry
fix two bugs found in testing.

Jan 23 2017, 9:20 PM
airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

updates after taking to Tom on irc, still might cause a flag day.

Jan 23 2017, 8:42 PM

Dec 28 2016

airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

This rewrites the patch, using a new triple OsType (needs renaming to something better), and adds an intrinsic to get access to the private buffer.

Dec 28 2016, 7:18 PM

Dec 15 2016

airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.
Dec 15 2016, 5:53 PM

Oct 10 2016

airlied updated the diff for D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

Updated diff to spill to a 64-bit buffer, pointed to by 64-bits in the buffer pointed to by SGPR0/1.

Oct 10 2016, 10:35 PM
airlied added a comment to D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

I think ABI changes should be encoded as part of the triple (a radv environment type?) rather than a subtarget feature. This is supposed to be just the pointer and not the full scratch resource descriptor?

Oct 10 2016, 12:49 AM

Oct 9 2016

airlied added a comment to D25428: AMDGPU add support for spilling to a user sgpr pointed buffers.

FYI: The current mesa side of this is here:
https://github.com/airlied/mesa/tree/radv-wip-spilling

Oct 9 2016, 11:55 PM
airlied retitled D25428: AMDGPU add support for spilling to a user sgpr pointed buffers from to AMDGPU add support for spilling to a user data SREG address..
Oct 9 2016, 11:53 PM