Page MenuHomePhabricator

sameerds (Sameer Sahasrabuddhe)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 17 2012, 11:25 AM (356 w, 5 d)

Working on ROCm at AMD

Recent Activity

Nov 16 2018

sameerds added a comment to D54516: [AMDGPU] Do not mark llvm.amdgcn.set.inactive as IntrNoMem.
In D54516#1301072, @tpr wrote:

EarlyCSE does seem to common up in this situation. And, if I disable that, I get GVN commoning it up.

Nov 16 2018, 9:26 AM

Sep 25 2018

sameerds committed rL342956: [AMDGPU] restore r342722 which was reverted with r342743.
[AMDGPU] restore r342722 which was reverted with r342743
Sep 25 2018, 2:41 AM

Sep 21 2018

sameerds committed rL342743: revert changes from r342722.
revert changes from r342722
Sep 21 2018, 9:33 AM
sameerds updated subscribers of D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

Sorry about that! I am reverting the change for now. Waiting for the tests
to pass locally with the change reverted.

Sep 21 2018, 9:02 AM
sameerds committed rL342722: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.
[AMDGPU] lower-switch in preISel as a workaround for legacy DA
Sep 21 2018, 4:28 AM
sameerds closed D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.
Sep 21 2018, 4:28 AM
sameerds updated the diff for D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

figured out how to fix the revision using "arc diff --edit"

Sep 21 2018, 4:06 AM

Sep 20 2018

sameerds added a comment to D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

I changed the first line in the change description to make it explicit that this is a workaround for DA, but it is not showing up here. I just used "arc diff master" ... how do I get the web interface to show the updated description?

Sep 20 2018, 9:44 AM
sameerds updated the diff for D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

addressed comments from @arsenm

Sep 20 2018, 9:39 AM

Sep 19 2018

sameerds added a comment to D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

Good to hear the new DA can handle this. We need lowered switches for the control flow lowering anyway, though, so we may as well do it this way.

The test case is extremely mysterious to me, though. What is it that's actually being tested here? From the description I understand it's the divergence of the phi in sw.epilog? I don't see how that relates to the CHECK lines.

Updated the test to better convey the intention. Does that help? Note that since this change is a backend-specific workaround, the test focuses on the symptom seen in the results of si-annotate-control-flow rather than checking the results of divergence analysis.

Sep 19 2018, 11:23 PM
sameerds added a reviewer for D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA: nhaehnle.
Sep 19 2018, 11:23 PM

Sep 18 2018

sameerds added a comment to D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

Good to hear the new DA can handle this. We need lowered switches for the control flow lowering anyway, though, so we may as well do it this way.

The test case is extremely mysterious to me, though. What is it that's actually being tested here? From the description I understand it's the divergence of the phi in sw.epilog? I don't see how that relates to the CHECK lines.

Sep 18 2018, 10:47 AM
sameerds added a comment to D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

This workaround will not longer be required with the new Divergence Analysis (https://reviews.llvm.org/D50433). I will add this test case as a unit test for the new implementation.

Sep 18 2018, 10:46 AM
sameerds updated the diff for D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.

improve the lit test

Sep 18 2018, 10:35 AM
sameerds created D52221: [AMDGPU] lower-switch in preISel as a workaround for legacy DA.
Sep 18 2018, 1:06 AM

Sep 2 2018

sameerds committed rL341295: Added publications for AHIR, an LLVM-based toolchain for high-level synthesis.
Added publications for AHIR, an LLVM-based toolchain for high-level synthesis
Sep 2 2018, 11:47 PM

Feb 24 2015

sameerds committed rL230464: Restores r228382, which was reverted in r228406..
Restores r228382, which was reverted in r228406.
Feb 24 2015, 9:50 PM

Feb 5 2015

sameerds committed rL228382: OpenCL: handle shift operator with vector operands.
OpenCL: handle shift operator with vector operands
Feb 5 2015, 9:46 PM

Feb 3 2015

sameerds committed rL228118: OpenCL: handle ternary operator when the condition is a vector.
OpenCL: handle ternary operator when the condition is a vector
Feb 3 2015, 10:40 PM

Jan 19 2015

sameerds committed rL226548: Introduce SPIR calling conventions..
Introduce SPIR calling conventions.
Jan 19 2015, 10:46 PM