Chromium Code Reviews
|
|
Chromium Code Reviews
Issue
1480553004:
[heap] Tweak evacuation candidate selection. (Closed)
|
Created:
5 years ago by Michael Lippautz Modified:
5 years ago CC:
v8-reviews_googlegroups.com Base URL:
https://chromium.googlesource.com/v8/v8.git@master Target Ref:
refs/pending/heads/master Project:
v8 Visibility:
Public. |
Description[heap] Tweak evacuation candidate selection.
Compaction selection now derives target fragmentation based on compaction speed
and a goal per page, if compaction speed has been traced. This way we avoid
overlong compaction of single pages.
This change could regress memory consumption if (a) the memory reducer does not
kick in, and (b) lots of medium fragmented pages would be compacted in our fast
path.
BUG=chromium:524425
LOG=N
Committed: https://crrev.com/97f1bac47e51d48c271e3144513ac0b90cad90ff
Cr-Commit-Position: refs/heads/master@{#32361}
Patch Set 1 : #Patch Set 2 : #
Total comments: 3
Patch Set 3 : Addressed comment #
Total comments: 4
Patch Set 4 : Factored out computing heuristics #Messages
Total messages: 27 (15 generated)
Description was changed from ========== [heap] Tweak compaction selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. BUG=chromium:524425 LOG=N ========== to ========== [heap] Tweak compaction selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ==========
Patchset #2 (id:20001) has been deleted
Patchset #2 (id:40001) has been deleted
Patchset #1 (id:1) has been deleted
Patchset #1 (id:60001) has been deleted
Description was changed from ========== [heap] Tweak compaction selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ========== to ========== [heap] Tweak evacuation candidate selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ==========
mlippautz@chromium.org changed reviewers: + hpayer@chromium.org, ulan@chromium.org
mlippautz@chromium.org changed reviewers: + hpayer@chromium.org, ulan@chromium.org
PTAL https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.... src/heap/mark-compact.cc:693: const int kTargetFragmentationPercent = 70; Increased the default from 50 to 70, making it less aggressive.
PTAL https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.... src/heap/mark-compact.cc:693: const int kTargetFragmentationPercent = 70; Increased the default from 50 to 70, making it less aggressive.
lgtm https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.... src/heap/mark-compact.cc:715: if (target_fragmentation_percent < 0) { Let's cap it from below by kTargetFragmentationPercentForReduceMemory or kTargetFragmentationPercent
https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/100001/src/heap/mark-compact.... src/heap/mark-compact.cc:715: if (target_fragmentation_percent < 0) { On 2015/11/26 14:15:59, ulan wrote: > Let's cap it from below by kTargetFragmentationPercentForReduceMemory or > kTargetFragmentationPercent Done. Lower cap is using kTargetFragmentationPercentForReduceMemory
The CQ bit was checked by mlippautz@chromium.org to run a CQ dry run
Dry run: CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1480553004/120001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1480553004/120001
The CQ bit was unchecked by commit-bot@chromium.org
Dry run: This issue passed the CQ dry run.
https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.... src/heap/mark-compact.cc:682: // them starting at page with the most free memory, adding them to the set ...starting with the page... https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.... src/heap/mark-compact.cc:689: This method is getting really long. Can we factor out this heuristic from the selection function?
Patchset #4 (id:140001) has been deleted
https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.cc File src/heap/mark-compact.cc (right): https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.... src/heap/mark-compact.cc:682: // them starting at page with the most free memory, adding them to the set On 2015/11/27 07:57:27, Hannes Payer wrote: > ...starting with the page... Done. https://codereview.chromium.org/1480553004/diff/120001/src/heap/mark-compact.... src/heap/mark-compact.cc:689: On 2015/11/27 07:57:27, Hannes Payer wrote: > This method is getting really long. Can we factor out this heuristic from the > selection function? Done.
lgtm
The CQ bit was checked by mlippautz@chromium.org
The patchset sent to the CQ was uploaded after l-g-t-m from ulan@chromium.org Link to the patchset: https://codereview.chromium.org/1480553004/#ps160001 (title: "Factored out computing heuristics")
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1480553004/160001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1480553004/160001
Message was sent while issue was closed.
Description was changed from ========== [heap] Tweak evacuation candidate selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ========== to ========== [heap] Tweak evacuation candidate selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ==========
Message was sent while issue was closed.
Committed patchset #4 (id:160001)
Message was sent while issue was closed.
Description was changed from ========== [heap] Tweak evacuation candidate selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N ========== to ========== [heap] Tweak evacuation candidate selection. Compaction selection now derives target fragmentation based on compaction speed and a goal per page, if compaction speed has been traced. This way we avoid overlong compaction of single pages. This change could regress memory consumption if (a) the memory reducer does not kick in, and (b) lots of medium fragmented pages would be compacted in our fast path. BUG=chromium:524425 LOG=N Committed: https://crrev.com/97f1bac47e51d48c271e3144513ac0b90cad90ff Cr-Commit-Position: refs/heads/master@{#32361} ==========
Message was sent while issue was closed.
Patchset 4 (id:??) landed as https://crrev.com/97f1bac47e51d48c271e3144513ac0b90cad90ff Cr-Commit-Position: refs/heads/master@{#32361} |
