AuroraMiddleware/v8
1
0
Fork 0
Code Issues 2 Pull Requests Releases Wiki Activity
15,744 Commits 1 Branch 0 Tags 839 MiB
20461079e9
Commit Graph

751 Commits

Author SHA1 Message Date
mvstanton@chromium.org
9e5757abee Revert "Reland of https://codereview.chromium.org/172523002/" 
This reverts commit r20516 due to a Sunspider performance issue.

R=ishell@chromium.org

Review URL: https://codereview.chromium.org/226233002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20521 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-04 16:18:59 +00:00
mvstanton@chromium.org
f9a8425cd2 Reland of https://codereview.chromium.org/172523002/ 
Fixed
1) Missing line in x64 port.
2) GcStress found a logic error in the IC miss handler.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/224903005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20516 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-04 13:57:53 +00:00
mvstanton@chromium.org
5c0d85163c Revert r20474 
This reverts commit r20474 due to an x64 release issue.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/223823002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20478 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-03 12:27:24 +00:00
mvstanton@chromium.org
c21cf2d334 Function calls are handled with a LoadIC and a FunctionCallStub. This works well, but we can create a platform to gather custom feedback by inventing a CallIC. CallIC takes a JSFunction as input, maintains UNINIT/MONO/MEGA state in a feedback slot, and engages in limited patching for the most efficient code. 
The change creates a degrade in NBody tests, but a follow-on CL addresses it by using this custom feedback platform to give a 45% improvement.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/172523002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20474 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-03 11:41:03 +00:00
jarin@chromium.org
0b53ed2d2b Check in Lithium that allocation size in Smi range. 
This is to avoid triggering an assertion from Smi::FromInt. The
generated code is unreachable, so it is not a real bug.

R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/221743005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20458 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-03 07:04:46 +00:00
svenpanne@chromium.org
4e40600485 Check that environments assigned via AssignEnvironment are actually used. 
Removed some temporary marker comments on the way.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/218403006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20430 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-02 11:30:13 +00:00
svenpanne@chromium.org
497487beec Consistently use a separate Lithium instruction for flooring division. 
Previously we tried to share some code on by a slightly confusing re-use
of LDivI for a (general) flooring division. Now we cleanly separate
concerns, just like for the rest of the division-like operations. Note
that ARM64 already did it this way.

If we really want to save some code, we can introduce some macro
assembler instructions and/or helper functions in the code generator in
a future CL, but we should really try to avoid being "clever" to save
just a few lines of trivial code. Effort != complexity. :-)

Renamed some related Lithium operands on the way for more consistency.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/212703002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20395 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-01 11:42:42 +00:00
svenpanne@chromium.org
4da732e321 Fixed power-of-2 predicates, excluding 0. 
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/219213003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20383 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-04-01 09:20:28 +00:00
dslomov@chromium.org
b3148d921e Fix PrepareKeyedOperand on arm. 
When additional_offset is specified, the 'key' operand can be negative
and still pass the bounds check. Therefore, when converting key from
Smi, arithmetic and not logical shift must be used.

R=verwaest@chromium.org
BUG=358057
LOG=Y

Review URL: https://codereview.chromium.org/219473002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20363 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-31 15:14:28 +00:00
verwaest@chromium.org
019e27d8db Reland and fix "Fix LoadFieldByIndex to take mutable heap-numbers into account."" 
BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/218663005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20358 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-31 14:21:04 +00:00
jochen@chromium.org
163044e7ba Revert 20348 - "Fix LoadFieldByIndex to take mutable heap-numbers into account." 
Reason for revert: crashes benchmarks/sunspider/string-fasta on ia32.debug

This also reverts r20350 and r20352

> Fix LoadFieldByIndex to take mutable heap-numbers into account.
>
> BUG=
> R=ishell@chromium.org
>
> Review URL: https://codereview.chromium.org/213213002

BUG=none
LOG=n
TBR=verwaest@chromium.org

Revert "Use sarq on x64"

This reverts commit e2a8ef9321345c6bc091054443bf2b9535ff6b1c.

Revert "Don't | int and bool"

This reverts commit c90d713d3a8ceba4fec41933a63beb6e50a3d7c0.

Review URL: https://codereview.chromium.org/219393002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20354 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-31 13:23:32 +00:00
verwaest@chromium.org
55a6318560 Fix LoadFieldByIndex to take mutable heap-numbers into account. 
BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/213213002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20348 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-31 11:59:29 +00:00
jarin@chromium.org
d65fe51ca0 Add missing lazy deopt point for the TransitionElementsKind instruction. 
R=mvstanton@chromium.org, yangguo@chromium.org
BUG=357105
TEST=test/mjsunit/regress/regress-357105.js
LOG=N

Review URL: https://codereview.chromium.org/216963002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20347 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-31 11:58:53 +00:00
yangguo@chromium.org
9be61ddb8a Hide some runtime functions. 
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/212163004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20285 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-26 15:51:48 +00:00
dslomov@chromium.org
f66af4feb4 Refactor optimized in hydrogen only runtime functions. 
This splits all runtime function into 3 categories:
1) RUNTIME: implemented in runtime and called from both full and optimized code.
2) RUNTIME_HIDDEN: implemented in runtime, never called directly from JS builtins.
3) INLINE: inlined in both full and optimized code
4) INLINE_OPTIMIZED: inlined in optimized code, implemented in runtime for full code.

R=yangguo@chromium.org, yannguo@chromium.org

Review URL: https://codereview.chromium.org/209353006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-25 14:26:55 +00:00
svenpanne@chromium.org
3081e1b730 Replaced unreachable code in DoStoreNamedField by an assertion. 
This is part of the effort to assign an environment if and only if we can deopt.

Removed some related dead code on the way.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/209873005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20194 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-24 13:16:23 +00:00
ulan@chromium.org
f20a9473f3 Ensure that lazy deopt sequence does not override calls. 
BUG=354433
LOG=N
TEST=mjsunit/regress/regress-354433.js
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/198463006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20155 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-21 11:02:15 +00:00
svenpanne@chromium.org
68237f6590 Implement flooring division by a constant via truncating division by a constant. 
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/204583002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20123 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-20 13:10:23 +00:00
mvstanton@chromium.org
5e4febb06e Array constructor expects AllocationSite or undefined as feedback. 
Redefine Array constructor to expect an AllocationSite in the feedback
register or undefined. This will make code simpler as we support
pretenuring feedback for all constructed objects.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/203173003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20064 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-19 11:51:30 +00:00
svenpanne@chromium.org
1a6f977d6a Tweaked FlooringDiv back to TruncatingDiv again. 
This is a first step towards getting the flooring division by a
constant working again (which will be a separate CL).

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/197873029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20048 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-19 07:01:08 +00:00
ulan@chromium.org
487ca9e384 Fix TransitionElementsKindStub to handle non-JSArray objects correctly. 
BUG=352982
LOG=N
TEST=mjsunit/regress/regress-352982.js
R=danno@chromium.org

Review URL: https://codereview.chromium.org/196343023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-18 13:29:29 +00:00
mvstanton@chromium.org
11df4b8815 Fix for issue 351261. 
This relands the following fix: "HAllocate should never generate
allocation code if the requested size does not fit into page. Regression
test included. (bug 347543)" along with additional fixes to KeyedStoreIC.

BUG=351261
LOG=N
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/200113002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19926 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-14 10:22:55 +00:00
bmeurer@chromium.org
4da15df008 Remove uses of CanBeNegative() in HMod. 
BUG=v8:3204
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/195793016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19878 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-13 08:17:44 +00:00
bmeurer@chromium.org
929313bd6b Remove uses of RangeCanInclude() in flooring division by power of 2. 
Drive-By-Fix: Improve ARM code generation for flooring division by
power of 2.

BUG=v8:3204
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/196653009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19877 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-13 07:58:58 +00:00
bmeurer@chromium.org
750f2d98f8 Fix uses of range analysis results in HChange. 
BUG=v8:3204
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/195023002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19872 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-13 06:11:52 +00:00
rmcilroy@chromium.org
7f622cffd4 Introduce FrameAndConstantPoolScope and ConstantPoolUnavailableScope. 
Adds FrameAndConstantPoolScope and ConstantPoolUnavailableScope to enable
scoped management of constant pool availability.  Also load constant pool
pointer when entering an internal frame scope.

R=rodolph.perfetta@arm.com, ulan@chromium.org

Review URL: https://codereview.chromium.org/190793002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19858 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-12 15:56:16 +00:00
svenpanne@chromium.org
d6f639bcb0 Handle flooring division in LCodeGen::DoDivByConstI on ARM, too. 
We should really split up the Lithium instruction, but this will be
done in some future cleanup CL.

Removed some "const"s for local variables on the way, they don't really
help us much and just clutter up the code.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/196603004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19850 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-12 14:06:26 +00:00
jkummerow@chromium.org
8a1812f252 Fix lazy deopt after tagged binary ops 
Also add policing code to ensure that optimized frames can in fact lazily deopt
at their respective current PC when we patch them for lazy bailout.

BUG=chromium:350434
LOG=y
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/194703008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19834 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-12 09:59:36 +00:00
rossberg@chromium.org
8e3f3cee9e Eliminate extended mode, and other modes clean-up 
- Merge LanguageMode and StrictModeFlag enums
- Make harmony-scoping depend only on strict mode
- Free some bits on the way
- Plus additional clean-up and renaming

R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/181543002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19800 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-11 14:41:22 +00:00
rossberg@chromium.org
3f702d4bf9 Mode clean-up pt 1: rename classic/non-strict mode to sloppy mode 
R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/177683002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19799 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-11 14:39:08 +00:00
bmeurer@chromium.org
4ac0876a8c Cleanup some of the range uses in ModI/DivI. 
BUG=v8:3204
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/191293013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-11 11:57:27 +00:00
m.m.capewell@googlemail.com
be8e2885d3 ARM: cleanning of InstanceOfKnownGlobal 
BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/189263011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19758 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-10 13:23:28 +00:00
bmeurer@chromium.org
bf86e624d4 Reland "Handle non-power-of-2 divisors in division-like operations". 
Fixed the flooring div bug and added a test case.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/191293012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19749 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-10 10:39:17 +00:00
yangguo@chromium.org
4f15fd2977 Reland "Introduce intrinsics for double values in Javascript." 
This relands r19704 with a fix to the test case.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/189823003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19723 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 14:58:41 +00:00
svenpanne@chromium.org
fa6d25a602 Revert "Handle non-power-of-2 divisors in division-like operations", "A64 tweaks for division-like operations." and "Windows build fix.". 
This reverts commit 19719, 19720 and 19721 because
mozilla/ecma/Date/15.9.3.1-1 fails (in release mode only?).

TBR=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/189963005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19722 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 13:11:56 +00:00
svenpanne@chromium.org
94c450fcb9 Handle non-power-of-2 divisors in division-like operations 
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/190383002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19719 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 11:44:04 +00:00
svenpanne@chromium.org
819315db4e Consistenly handle power-of-2 divisors in division-like operations 
Lithium currently supports 3 division-like operations on integral operands: "Normal" division (rounding towards zero), flooring division (rounding towards -Infinity) and modulus calculation (the counterpart for the "normal" division). For divisors which are a power of 2, one can efficiently use some bit fiddling to avoid the actual division for such operations. This CL cleanly splits off these operations into separate Lithium instructions, making the code much more maintainable and more consistent across platforms.

There are 2 basic variations of these bit fiddling algorithms: One involving branches and a seemingly more clever one without branches. Choosing between the two is not as easy as it seems: Benchmarks (and probably real-world) programs seem to favor positive dividends, registers and shifting units are sometimes scarce resources, and branch prediction is quite good in modern processors. Therefore only the "normal" division by a power of 2 is implemented in a branch-free manner, this seems to be the best approach in practice. If this turns out to be wrong, we can easily and locally change this.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/175143002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19715 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 10:36:28 +00:00
yangguo@chromium.org
143902bebf Revert "Introduce intrinsics for double values in Javascript." 
This reverts r19704.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/189533008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19710 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 09:49:28 +00:00
mvstanton@chromium.org
1d3652ebe6 Symbols for type cells. We can make more efficient code to check against type cells in the future if we use symbols, guaranteed not to conflict with user code. Currently, the "symbols" are the hole and undefined. Undefined may come in from the outside. 
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/181283003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19706 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 09:10:18 +00:00
yangguo@chromium.org
2aefde4443 Introduce intrinsics for double values in Javascript. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/178583006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19704 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-07 09:05:10 +00:00
mvstanton@chromium.org
6115a006fd Bugfix for 349874: we incorrectly believe we saw a growing store 
When we set an out of bounds array index, the index might be so large that
it causes the array to go to dictionary mode. It's better to avoid
"learning" that this was a growing store in that case.

This fix also partially reverts a fix for bug 347543, as this fix is
comprehensive and satisfies that repro case as well (partial revert of
v19591).

BUG=349874
LOG=N
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/188643002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19691 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-06 13:07:51 +00:00
jarin@chromium.org
7ac668f753 Deoptimization fix for HPushArgument. 
HPushArgument should never be used in a simulation environment
because the slot addresses for the arguments can be off (e.g.,
due to on-stack arguments object of an inlined caller).

R=mstarzinger@chromium.org
BUG=v8:3183
LOG=N

Review URL: https://codereview.chromium.org/178193026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19675 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-05 12:45:46 +00:00
verwaest@chromium.org
9b8021ad29 Remove all uses of field-tracking flags that do not make decisions but are subject to existing information. 
BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/173963002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19650 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-04 12:48:17 +00:00
bmeurer@chromium.org
e16cc0acaf Push safepoint registers in deferred number-to-i/u only on-demand. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/181053005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19649 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-04 12:45:00 +00:00
ulan@chromium.org
b9e0b87a5a Clear optimized code cache in shared function info when code gets deoptimized. 
This adds a pointer to the shared function info into deoptimization data of an optimized code. Whenever the code is deoptimized, it clears the cache in the shared function info.

This fixes the problem when the optimized function dies in new space GC before the code is deoptimized due to code dependency and before the optimized code cache is cleared in old space GC (see mjsunit/regress/regress-343609.js).

This partially reverts r19603 because we need to be able to evict specific code from the optimized code cache.

BUG=343609
LOG=Y
TEST=mjsunit/regress/regress-343609.js
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/184923002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-03-03 11:11:39 +00:00
ishell@chromium.org
2ab83cf192 HAllocate should never generate allocation code if the requested size does not fit into page. Regression test included. 
BUG=347543
LOG=N
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/180803005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19591 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-02-27 17:33:25 +00:00
ulan@chromium.org
db60894687 Clean up weak objects in optimized code. 
This prepares for weak objects in IC stubs.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/172233003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19489 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-02-19 14:03:48 +00:00
yangguo@chromium.org
139134acc2 Harmony: optimize Math.clz32. 
R=svenpanne@chromium.org
BUG=v8:2938
LOG=N

Review URL: https://codereview.chromium.org/172133003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19487 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-02-19 13:51:49 +00:00
svenpanne@chromium.org
dbce27047e Fixed and improved code for integral division. Fixed and extended tests. 
Arithmetic right shifting is *not* division in two's complement
representation, only in one's complement. So we convert to one's
complement, shift, and go back to two's complement. By permutating the
last steps, one can get efficient branch-free code. This insight comes
from the paleozoic era of computer science, see the paper from 1976:

   Guy Lewis Steele Jr.: "Arithmetic Shifting Considered Harmful"
   ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-378.pdf

This results in better and more correct code than our previous
"neg/shift/neg" dance.

LOG=y
BUG=v8:3151
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/166793002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19434 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-02-18 10:45:27 +00:00
vegorov@chromium.org
8f170a66e7 Improve positions tracking inside the HGraphBuilder. 
Instead of tracking simple absolute offset from the start of the script like other places do, track a pair of (inlining id, offset from the start of inlined function).

This enables us to pinpoint with inlining path an instruction came from. Previously in multi-script environments we emitted positions that made very little sense because inside a single optimized function they would point to different scripts without a way to distinguish them.

Start dumping the source of every inlined function to make possible IR viewing tools with integrated source views as there was previously no way to acquire this information from IR dumps. We also dump source position at which each inlining occured.

Tracked positions are written into hydrogen.cfg as pos:<inlining-id>_<offset>.

Flag --emit-opt-code-positions is renamed by this change into --hydrogen-track-positions to better convey it's meaning.

In addition this change assigned global unique identifier to each optimization performed inside isolate. This allows to precisely match compilation artifacts (e.g. IR and disassembly) and deoptimizations.

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/140683011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19360 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-02-13 16:09:28 +00:00