This is an improved version of my earlier change r5970. It avoids degrading the
non-optimized code.
Initially we emit a conditional branch that is either always- or never-taken
after a smi-check (depending on whether we test for smi for for non-smi)
Since test-eax always sets the carry-flag to 0 we use jump-if-carry and
jump-if-not-carry.
The first invocation of the stub patches a jc with a jz and
jnc with a jnz-instruction so that the code looks exactly as it was
without patching. The only difference is the test- or nop-instruction
after the IC-call.
Review URL: http://codereview.chromium.org/5763004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6030 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
In the case of inlined smi code in non-optimzied code we could not
distinguish between the smi-only case and the case that the operation was
never executed.
With this change the first execution of a binary operation always jumps
to the stub which in turn patches the smi-check into the correct
conditional branch, so that we benefit from inlined smi code after the
first invocation.
A nop instruction after the call to the BinaryOpIC indicates that no
smi code was inlined. A "test eax" instruction says that there was smi
code inlined and encodes the delta to the patch site and the condition
code of the branch at the patch site to restore the original jump.
Review URL: http://codereview.chromium.org/5714001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5970 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Contextual load requires only a map check followed by a cell hole
check so we can generate pretty compact code for that. The fact that
we have inlined code is marked by mov ecx, offset instruction after
the IC call. Inlining is only enabled inside loops and in non-builtin
functions.
The generated code size increase is about 3%. This descreased the
pc-to-code cache hit rate in some of the benchmarks that trigger
GC. To compensate we now have 4 times as much entries in the cache.
Review URL: http://codereview.chromium.org/3402014
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5497 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Object model changes
----------------------------------------
New fixed_cow_array_map is used for the elements array of a JSObject
to mark it as COW. The JSObject's map and other fields are not
affected. The JSObject's map still has the "fast elements" bit set. It
means we can do only the receiver map check in keyed loads and the
receiver and the elements map checks in keyed stores. So introducing
COW arrays doesn't hurt performance of these operations. But note that
the elements map check is necessary in all mutating operations because
the "has fast elements" bit now means "has fast elements for reading".
EnsureWritableFastElements can be used in runtime functions to perform
the necessary lazy copying.
Generated code changes
----------------------------------------
Generic keyed load is updated to only do the receiver map check (this
could have been done earlier). FastCloneShallowArrayStub now has two
modes: clone elements and use COW elements. AssertFastElements macro
is added to check the elements when necessary. The custom call IC
generators for Array.prototype.{push,pop} are updated to avoid going
to the slow case (and patching the IC) when calling the builtin should
work.
COW enablement
----------------------------------------
Currently we only put shallow and simple literal arrays in the COW
mode. This is done by the parser.
Review URL: http://codereview.chromium.org/3144002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The IC stub is completely generic, so there will only be one such stub
in the system.
Added a new overloaded version of the macro assembler RecordWrite
method for cases where we have the address we store to computed up
front.
Review URL: http://codereview.chromium.org/2804029
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4991 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Similar or duplicate checks are scattered around the code before doing the dictionary load.
Also the entire branch in GenerateCallNormal that handles global/builtin receiver is
guaranteed to bail out from GenerateDictionaryLoad, so there is no point in generating it at all.
The purpose of the patch is:
- making C++ code more compact and transparent,
- not generating dead code.
There is a tiny performance gain. The patch is ia32 only for now.
Please tell me if I am missing anything.
Review URL: http://codereview.chromium.org/2801007
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4926 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
These string methods can be composed from two basic blocks: charCodeAt
and fromCharCode, both of which have fast cases for certain types of
inputs. In this patch these two blocks are refactored to allow
generating the fast cases without having to jump around the slow
cases. In the slow cases since they can now be invoked both from
inline runtime functions and from IC stubs we either have to
save/restore state of the current frame or enter/leave a new internal
frame. This is handled by new RuntimeCallHelper interface. Its
implementation for virtual frame is based on FrameRegisterState class
extracted from DeferredCode class.
Review URL: http://codereview.chromium.org/2087009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4733 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
- New сardmarking write barrier handles large objects and normal objects in a similar fashion (no more additional space for pointer tracking is required, no conditional branches in WB code).
- Changes to enable oldspaces iteration without maps decoding:
-- layout change for FixedArrays: length is stored as a smis (initial patch by
Kevin Millikin)
-- layout change for SharedFunctionInfo: integer fields are stored as smi on
arm, ia32 and rearranged on x64.
-- layout change for String: meaning of LSB bit is fliped (1 now means hash not
computed); on x64 padding is added.
-- layout of maps is _not_ changed. Map space is currently iterated in a special
way.
Review URL: http://codereview.chromium.org/2144006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4715 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-- layout change for FixedArrays: length is stored as a smis (initial patch by Kevin Millikin)
-- layout change for SharedFunctionInfo: integer fields are stored as smi on arm, ia32 and rearranged on x64.
-- layout change for String: meaning of LSB bit is fliped (1 now means hash not computed); on x64 padding is added.
-- layout of maps is _not_ changed. Map space is currently iterated in a special way.
- Cardmarking write barrier. New barrier handles large objects and normal objects in a similar fashion (no more additional space for pointer tracking is required, no conditional branches in WB code).
Review URL: http://codereview.chromium.org/2101002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4685 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Chromium build.
v8.gyp no longer sets any V8_TARGET_ARCH_* macro on the Mac. Instead, the
proper V8_TARGET_ARCH_* macro will be set by src/globals.h in the same way as
the V8_HOST_ARCH_* macro when it detects that no target macro is currently
defined. The Mac build will attempt to compile all ia32 and x86_64 .cc files.
#ifdef guards in each of these target-specific source files prevent their
compilation when the associated target is not selected. For completeness,
these #ifdef guards are also provided for the arm and mips .cc files.
BUG=706
TEST=x86_64 Mac GYP/Xcode-based Chromium build (still depends on other changes)
Review URL: http://codereview.chromium.org/2133003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4666 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
(Fixed handling of out-of-bounds keys.)
String keyed load used to call STRING_CHAR_AT builtin that performs
two steps (get a char code, construct a one-char string from the
code), both of which have fast cases implemented as inline runtime
functions. In this chage most of the code from these functions is
extracted to a set of common generator functions in StringStubBase and
the fast cases are grouped together in the IC code.
Review URL: http://codereview.chromium.org/1582041
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4450 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
String keyed load used to call STRING_CHAR_AT builtin that performs
two steps (get a char code, construct a one-char string from the
code), both of which have fast cases implemented as inline runtime
functions. In this chage most of the code from these functions is
extracted to a set of common generator functions in StringStubBase and
the fast cases are grouped together in the IC code.
Review URL: http://codereview.chromium.org/1539039
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4444 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
With my previous change to limit memory for object literals, we get more slow-case elements and this makes up for the slowdown when loading from those slow-case elements.
The most complicated part here is the computation of the integer hash code. We might want to simplify the integer hash function.
Review URL: http://codereview.chromium.org/857003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4109 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Instead of going through a runtime function for keyed loads
on strings we invoke a separate specialized stub that
assumes string as receiver type and the key to be a number.
The stub calls a JS builtin function to return the corresponding one-character string.
Review URL: http://codereview.chromium.org/521041
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3556 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The different length string types was used to encode the string length and the hash in one field. This is now split into two fields one for length and one for hash. The hash field still encodes the array index of the string if it has one. If an array index is encoded in the hash field the string length is added to the top bits of the hash field to avoid a hash value of zero.
On 32-bit this causes an additional 4 bytes to be used for all string objects. On 64-bit this will be half on average dur to pointer alignment.
Review URL: http://codereview.chromium.org/436001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
specification under development. This is a follow-on CL to
http://codereview.chromium.org/293023 .
Based on review feedback, defined the behavior of storing NaN and
+/-Infinity into external arrays of integer types as storing 0. Added
test cases. Added fucomi instruction to assembler. Fixed bug in
KeyedLoadIC::GenerateExternalArray when allocation of HeapNumber
failed. Fixed bug in encoding of 16-bit immediate arithmetic
instructions in 64-bit port.
Removed raising of exceptions for negative array indices passed to
external arrays and associated tests. Based on current discussion in
WebGL working group, will probably end up removing the exception
throwing altogether.
Review URL: http://codereview.chromium.org/294022
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3113 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
specification under development. The optimizations are patterned after
those previously done for CanvasPixelArray. This CL adds all of the
necessary framework but continues to use the generic KeyedLoadIC and
KeyedStoreIC code, to create a baseline for benchmarking purposes. The
next CL will add the optimized ICs to ic-ia32.cc and ic-x64.cc.
These new CanvasArray types have different semantics than
CanvasPixelArray; out-of-range values are clamped via C cast
semantics, which is cheaper than the clamping behavior specified by
CanvasPixelArray. Out-of-range indices raise exceptions instead of
being silently ignored.
As part of this work, pulled FloatingPointHelper::AllocateHeapNumber
up to MacroAssembler on ia32 and x64 platforms. Slightly refactored
KeyedLoadIC and KeyedStoreIC. Fixed encoding for fistp_d on x64 and
added a few more instructions that are needed for the new ICs. The
test cases in test-api.cc have been verified by hand to exercise all
of the generated code paths in the forthcoming specialized ICs.
Review URL: http://codereview.chromium.org/293023
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3096 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
called from within a loop or not. In the past we lost the
information if a call site went megamorphic before a lazily
compiled callee was called for the first time. Now we track
that correctly (this is an issue that affects richards).
We still don't manage to track the in-loop state through a
constructor call, since constructor calls use LoadICs instead
of CallICs. This issue affects delta-blue. So in this patch
we assume that lazy compilations that don't happen through a
CallIC happen from inside a loop. I have an idea to fix this
but this patch is big enough already.
With our improved tracking of in-loop state I have switched
off the inlining of in-object loads for code that is not in
a loop. This benefits compile speed. One issue is that
eagerly compiled code now doesn't get the in-object loads
inlined. We need to eagerly compile less code to fix this.
Review URL: http://codereview.chromium.org/115744
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2046 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This issue was raised by Brett Wilson while reviewing my changelist for readability. Craig Silverstein (one of C++ SG maintainers) confirmed that we should declare one namespace per line. Our way of namespaces closing seems not violating style guides (there is no clear agreement on it), so I left it intact.
Review URL: http://codereview.chromium.org/115756
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00