Issue 8160043: runtime: cache most immediately-computed hash value

Keyboard Shortcuts

	File
u :	up to issue
m :	publish + mail comments
M :	edit review message
j / k :	jump to file after / before current file
J / K :	jump to next file with a comment after / before current file
	Side-by-side diff
i :	toggle intra-line diffs
e :	expand all comments
c :	collapse all comments
s :	toggle showing all comments
n / p :	next / previous diff chunk or comment
N / P :	next / previous comment
<Up> / <Down> :	next / previous line
<Enter> :	respond to / edit current comment
d :	mark current comment as done

	Issue
u :	up to list of issues
m :	publish + mail comments
j / k :	jump to patch after / before current patch
o / <Enter> :	open current patch in side-by-side view
i :	open current patch in unified diff view

	Issue List
j / k :	jump to issue after / before current issue
o / <Enter> :	open current issue
# :	close issue

	Comment/message editing
<Ctrl> + s or <Ctrl> + Enter :	save comment
<Esc> :	cancel edit

Issue 8160043: runtime: cache most immediately-computed hash value (Closed)

Can't Edit
Can't Publish+Mail
Start Review

Created:
11 years ago by bradfitz

Modified:
10 years, 10 months ago

Reviewers:
dvyukov

Visibility:
Public.

Description

runtime: cache most immediately-computed hash value Proof of concept. NOT FOR SUBMISSION. Not enough benchmarks, but to show that it's working on string keys at least, and slowing things down when it's not triggering. benchmark old ns/op new ns/op delta BenchmarkNewEmptyMap 144 142 -1.39% BenchmarkHashStringSpeed 42 46 +10.21% BenchmarkHashInt32Speed 34 31 -8.45% BenchmarkHashInt64Speed 31 32 +2.52% BenchmarkHashStringArraySpeed 112 126 +12.50% BenchmarkMegMap 298981 28 -99.99% BenchmarkMegOneMap 24 24 +0.40% BenchmarkMegEmptyMap 4 4 +0.44% BenchmarkSmallStrMap 23 23 +0.00% BenchmarkIntMap 17 17 +0.56% BenchmarkRepeatedLookupStrMapKey32 45 35 -20.22% BenchmarkRepeatedLookupStrMapKey1M 306886 154843 -49.54% For Issue 5147

Patch Set 1 #

Patch Set 2 : diff -r 8f36f2ef721b https://go.googlecode.com/hg/ #

Patch Set 3 : diff -r 8f36f2ef721b https://go.googlecode.com/hg/ #

Total comments: 4

Patch Set 4 : diff -r a3717460e37c https://go.googlecode.com/hg/ #

Created: 11 years ago

Download [raw] [tar.bz2]

	Unified diffs	Side-by-side diffs	Delta from patch set	Stats (+109 lines, -10 lines)			Patch
M	src/pkg/runtime/hashmap.c	View	1 2	8 chunks	+72 lines, -6 lines	0 comments	Download
M	src/pkg/runtime/hashmap_fast.c	View	1	2 chunks	+26 lines, -4 lines	0 comments	Download
M	src/pkg/runtime/proc.c	View	1 2 3	6 chunks	+6 lines, -0 lines	0 comments	Download
M	src/pkg/runtime/runtime.h	View	1	1 chunk	+5 lines, -0 lines	0 comments	Download

Messages

Total messages: 2

Expand All Messages | Collapse All Messages

dvyukov

https://codereview.appspot.com/8160043/diff/3001/src/pkg/runtime/proc.c File src/pkg/runtime/proc.c (right): https://codereview.appspot.com/8160043/diff/3001/src/pkg/runtime/proc.c#newcode259 src/pkg/runtime/proc.c:259: mp->lasthash = 0; no need to initialize to zeroes, ...

11 years ago (2013-03-29 17:21:00 UTC) #1

bradfitz

11 years ago (2013-03-29 17:36:49 UTC) #2

Thanks for the review! (I'm new here.) PTAL

https://codereview.appspot.com/8160043/diff/3001/src/pkg/runtime/proc.c
File src/pkg/runtime/proc.c (right):

https://codereview.appspot.com/8160043/diff/3001/src/pkg/runtime/proc.c#newco...
src/pkg/runtime/proc.c:259: mp->lasthash = 0;
On 2013/03/29 17:21:00, dvyukov wrote:
> no need to initialize to zeroes, mp is initially all zeroes

Done. Thanks! I wasn't sure.

https://codereview.appspot.com/8160043/diff/3001/src/pkg/runtime/proc.c#newco...
src/pkg/runtime/proc.c:976: m->lastlen = 0; // clear the hash cache, in case a
GC occurred earlier
On 2013/03/29 17:21:00, dvyukov wrote:
> it needs to be in runtime.park/goexit/gosched/entersyscall/entersyscallblock
> otherwise (1) GC will see it and (2) hashmap/key can be garbage collected
while
> g is in syscall

Done.

Expand All Messages | Collapse All Messages