Chromium Code Reviews
|
|
Chromium Code Reviews
Issue
1677023002:
[turbofan] ARM: Improve AND instruction selection (Closed)
|
Created:
4 years, 10 months ago by martyn.capewell Modified:
4 years, 10 months ago CC:
v8-reviews_googlegroups.com, v8-mips-ports_googlegroups.com, v8-ppc-ports_googlegroups.com Base URL:
https://chromium.googlesource.com/v8/v8.git@master Target Ref:
refs/pending/heads/master Project:
v8 Visibility:
Public. |
Description[turbofan] ARM: Improve AND instruction selection
Improve instruction selector for mask and shift operations by using cheaper
instructions where possible, in preference to UBFX.
BUG=
Committed: https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845
Cr-Commit-Position: refs/heads/master@{#33843}
Patch Set 1 #
Messages
Total messages: 20 (7 generated)
Description was changed from ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= ========== to ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= ==========
martyn.capewell@arm.com changed reviewers: + bmeurer@chromium.org, danno@chromium.org, jarin@chromium.org
Nice, thanks! LGTM.
The CQ bit was checked by martyn.capewell@arm.com
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1677023002/1 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1677023002/1
Message was sent while issue was closed.
Description was changed from ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= ========== to ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= ==========
Message was sent while issue was closed.
Committed patchset #1 (id:1)
Message was sent while issue was closed.
Description was changed from ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= ========== to ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= Committed: https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845 Cr-Commit-Position: refs/heads/master@{#33843} ==========
Message was sent while issue was closed.
Patchset 1 (id:??) landed as https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845 Cr-Commit-Position: refs/heads/master@{#33843}
Message was sent while issue was closed.
On 2016/02/09 11:14:34, commit-bot: I haz the power wrote: > Patchset 1 (id:??) landed as > https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845 > Cr-Commit-Position: refs/heads/master@{#33843} Looks like this turned flaky after this CL: https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm%20-%20ar...
Message was sent while issue was closed.
Can it also be that it causes these timeouts on our chrubuntu chromebooks in gc stress mode? https://build.chromium.org/p/client.v8/builders/V8%20Arm%20GC%20Stress/builds...
Message was sent while issue was closed.
A revert of this CL (patchset #1 id:1) has been created in https://codereview.chromium.org/1681953003/ by martyn.capewell@arm.com. The reason for reverting is: Possibly causing Mozilla test failures - will investigate..
Message was sent while issue was closed.
Hmm - maybe this got generally flaky. The mozilla test also fails on another bot after the revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm%20-%20si...
Message was sent while issue was closed.
On 2016/02/10 11:30:58, Michael Achenbach wrote: > Hmm - maybe this got generally flaky. The mozilla test also fails on another bot > after the revert: > https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm%20-%20si... I think so. I've run a simulator build with and without patch 1000 times, and get similar results: cat log-unpatched.txt | grep -Po "(PASSED|FAILED)" | sort | uniq -c 899 FAILED 101 PASSED cat log-patched.txt | grep -Po "(PASSED|FAILED)" | sort | uniq -c 901 FAILED 99 PASSED So, I don't think this patch caused it. I'm still checking the Chromebook timeout.
Message was sent while issue was closed.
I've rerun the timing out test you mentioned on our Cortex-A15 Chromebook a number of times and get similar test durations with and without this patch. Though it can vary by ~30% irrespective of patch. So, I don't think either of the bot failures came from this patch. If you agree, I'll try relanding it.
Message was sent while issue was closed.
Description was changed from ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= Committed: https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845 Cr-Commit-Position: refs/heads/master@{#33843} ========== to ========== [turbofan] ARM: Improve AND instruction selection Improve instruction selector for mask and shift operations by using cheaper instructions where possible, in preference to UBFX. BUG= Committed: https://crrev.com/53d9c12977f07f55b6f2a72128b8d02c4c857845 Cr-Commit-Position: refs/heads/master@{#33843} ==========
Message was sent while issue was closed.
martyn.capewell@arm.com changed reviewers: + machenbach@chromium.org
Message was sent while issue was closed.
Adding Michael so he can see my comments.
Message was sent while issue was closed.
On 2016/02/12 10:47:17, martyn.capewell wrote: > Adding Michael so he can see my comments. Fair enough, try to reland. If it fails more again, I'll disable the tests and file a bug... |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
