Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(500)

Issue 10966031: Lowered kMaxVirtualRegisters (fixes v8 2139 and chrome 123822 and 128252). (Closed)

Created:
8 years, 3 months ago by Massi
Modified:
8 years, 2 months ago
CC:
v8-team_google.com
Visibility:
Public.

Description

Lowered kMaxVirtualRegisters (fixes v8 2139 and chrome 123822 and 128252). BUG=128252 Committed: https://code.google.com/p/v8/source/detail?r=12613

Patch Set 1 #

Total comments: 1

Patch Set 2 : Changed shift to 14. #

Patch Set 3 : Changed shift to 15. #

Patch Set 4 : Changed shift to 16. #

Patch Set 5 : Settled on 15 bits, redistributed bits to kFixedIndexWidth. #

Unified diffs Side-by-side diffs Delta from patch set Stats (+5 lines, -3 lines) Patch
M src/lithium.h View 1 2 3 4 2 chunks +5 lines, -3 lines 0 comments Download

Messages

Total messages: 6 (0 generated)
Massi
8 years, 3 months ago (2012-09-21 11:18:27 UTC) #1
Vyacheslav Egorov (Google)
dbc [I would say that big functions should be sent to background thread for compilation ...
8 years, 3 months ago (2012-09-21 11:42:08 UTC) #2
Massi
I totally agree on the separate thread approach, apart from the fact that we also ...
8 years, 3 months ago (2012-09-21 13:53:21 UTC) #3
Jakob Kummerow
lgtm
8 years, 3 months ago (2012-09-21 13:58:14 UTC) #4
Massi
To clarify, with "I tried multiple benchmarks" I meant these golem4 runs: http://70.32.156.88:8080/Comparison#targetA%3Dv8%3BmachineTypeA%3Dlinux-ia32%3BrevisionA%3D12548%3BpatchA%3Dmmassi-Fix-128252%3BtargetB%3Dv8%3BmachineTypeB%3Dlinux-ia32%3BrevisionB%3D12548%3BpatchB%3DNone (the 4 ...
8 years, 3 months ago (2012-09-21 14:01:23 UTC) #5
Sven Panne
8 years, 2 months ago (2012-10-01 09:40:41 UTC) #6
Just another DBC (late, but it's the 1st day after my vacation :-) :

When I looked into Alon Zakai's Emscripten version of Box2D (taken from his
blog), I found out that we actually need to *increase* these limits a lot,
otherwise we suck heavily because we don't crankshaft the hot functions. So this
CL is just hiding a symptom and makes things worse for other benchmarks.

I am not sure if we really need the background compilation approach for this
(although we should really try it), just deopting less  and speeding up
Crankshaft itself might be a viable direction, too...

Powered by Google App Engine
This is Rietveld 408576698