Issue 23684056: NumberUntagD is faster when untagging in a temp register

Keyboard Shortcuts

	File
u :	up to issue
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

	Issue
u :	up to list of issues
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

Issue 23684056: NumberUntagD is faster when untagging in a temp register (Closed)

Created:
7 years, 3 months ago by oliv

Modified:
7 years, 3 months ago

Reviewers:
ulan

CC:
v8-dev

Base URL:
https://v8.googlecode.com/svn/branches/bleeding_edge

Visibility:
Public.

More Reviews

Description

NumberUntagD is faster when untagging in a temp register BUG= https://code.google.com/p/v8/source/detail?r=16690

Patch Set 1 #

Total comments: 1

Created: 7 years, 3 months ago

Download [raw] [tar.bz2]

		Unified diffs	Side-by-side diffs	Delta from patch set	Stats (+14 lines, -13 lines)			Patch
	M	src/ia32/lithium-codegen-ia32.cc	View		3 chunks	+13 lines, -10 lines	0 comments	Download
	M	src/ia32/lithium-ia32.cc	View		1 chunk	+1 line, -3 lines	1 comment	Download

Messages

Total messages: 3 (0 generated)

Expand Messages | Collapse Messages

ulan

LGTM if it is faster on benchmarks. https://codereview.chromium.org/23684056/diff/1/src/ia32/lithium-ia32.cc File src/ia32/lithium-ia32.cc (right): https://codereview.chromium.org/23684056/diff/1/src/ia32/lithium-ia32.cc#newcode1936 src/ia32/lithium-ia32.cc:1936: LOperand* temp ...

7 years, 3 months ago (2013-09-12 08:35:04 UTC) #2

oliv

7 years, 3 months ago (2013-09-12 09:00:04 UTC) #3

On 2013/09/12 08:35:04, ulan wrote:
> LGTM if it is faster on benchmarks.
> 
> https://codereview.chromium.org/23684056/diff/1/src/ia32/lithium-ia32.cc
> File src/ia32/lithium-ia32.cc (right):
> 
>
https://codereview.chromium.org/23684056/diff/1/src/ia32/lithium-ia32.cc#newc...
> src/ia32/lithium-ia32.cc:1936: LOperand* temp = TempRegister();
> Doesn't this increase register pressure?

Yes it does, it's a tradeoff, on the other hand the cpu does not have to keep
the untagged value live.

since the untagD instruction usually occurs around other arithmetic
instructions, its a good assumption that the alu is saturated and it pays off to
avoid the re-tagging. in other situations it will not pay off...

Expand Messages | Collapse Messages