Bug 236662

Summary: [libpas] Use segmented vector for TLC layout list
Product: WebKit Reporter: Yusuke Suzuki <ysuzuki>
Component: New BugsAssignee: Yusuke Suzuki <ysuzuki>
Status: RESOLVED FIXED    
Severity: Normal CC: annulen, ews-watchlist, gyuyoung.kim, ryuan.choi, saam, sergio, webkit-bug-importer
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch
ews-feeder: commit-queue-
Patch
none
Patch
none
Patch
none
Patch
none
Patch
none
Patch
none
Patch
none
Patch
none
Patch saam: review+

Description Yusuke Suzuki 2022-02-15 12:59:55 PST 
[libpas] Use segmented vector for TLC layout list
Comment 1 Yusuke Suzuki 2022-02-15 13:00:14 PST 
Created attachment 452077 [details]
Patch
Comment 2 Yusuke Suzuki 2022-02-15 15:26:38 PST 
Created attachment 452102 [details]
Patch
Comment 3 Yusuke Suzuki 2022-02-15 17:33:21 PST 
Created attachment 452117 [details]
Patch
Comment 4 Yusuke Suzuki 2022-02-15 20:26:45 PST 
Created attachment 452129 [details]
Patch
Comment 5 Yusuke Suzuki 2022-02-15 21:02:16 PST 
Created attachment 452138 [details]
Patch
Comment 6 Yusuke Suzuki 2022-02-16 01:13:07 PST 
Created attachment 452158 [details]
Patch
Comment 7 Yusuke Suzuki 2022-02-16 01:27:56 PST 
Created attachment 452159 [details]
Patch
Comment 8 Yusuke Suzuki 2022-02-16 14:05:42 PST 
Created attachment 452247 [details]
Patch
Comment 9 Yusuke Suzuki 2022-02-16 17:39:08 PST 
Created attachment 452278 [details]
Patch
Comment 10 Yusuke Suzuki 2022-02-16 17:59:01 PST 
Created attachment 452282 [details]
Patch
Comment 11 Saam Barati 2022-02-16 18:49:32 PST 
Comment on attachment 452282 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=452282&action=review

Nice! r=me

> Source/bmalloc/ChangeLog:9
> +        linked-list of that to accelerate iteration performance. It does not affect on speed,

What do you mean by "It does not affect on speed"? Isn't the idea to make this faster, as you say below

> Source/bmalloc/ChangeLog:13
> +        well without taking a lock in most times which is blocked by scavenger, making scavenger
> +        faster is good for energy efficiency.

Maybe instead of speed, you can say "throughput of non scavenger threads"? Since this is speeding up the scavenger thread, but speeding up the scavenger thread doesn't speed up allocator threads.
Comment 12 Yusuke Suzuki 2022-02-16 23:41:56 PST 
OK, fixed, thanks!
Comment 13 Yusuke Suzuki 2022-02-16 23:42:38 PST 
Committed r290000 (247386@trunk): <https://commits.webkit.org/247386@trunk>
Comment 14 Radar WebKit Bug Importer 2022-02-16 23:43:17 PST 
<rdar://problem/89071706>
Comment 15 Yusuke Suzuki 2022-02-17 08:56:18 PST 
Committed r290024 (247404@trunk): <https://commits.webkit.org/247404@trunk>