- Increased size of number string cache....

- Increased size of number string cache.
- Change the instruction order for inlined allocation.


Committed: http://code.google.com/p/v8/source/detail?r=3514

[Kasper Lund, 2009-12-22 11:25:58]

LGTM.

http://codereview.chromium.org/501170/diff/1/2
File src/heap.cc (right):

http://codereview.chromium.org/501170/diff/1/2#newcode1613
src/heap.cc:1613: static inline int double_get_hash(double d) {
This should really be DoubleGetHash (the same goes for the smi variant).

http://codereview.chromium.org/501170/diff/1/2#newcode1616
src/heap.cc:1616: return (((value >> 16) ^ value))
How about adding another helper function so you can share the ((value >> 16) ^
value)) & (Heap::kNumberStrnigCacheSize - 1) code? (and maybe assert that
Heap::kNumberStringCacheSize is a power of two)

http://codereview.chromium.org/501170/diff/1/4
File src/heap.h (right):

http://codereview.chromium.org/501170/diff/1/4#newcode824
src/heap.h:824: static const int kNumberStringCacheSize = 16384;
How about 16 * KB or 16 * 1024?

[sra1, 2009-12-23 00:56:02]

http://codereview.chromium.org/501170/diff/1/3
File src/ia32/macro-assembler-ia32.cc (right):

http://codereview.chromium.org/501170/diff/1/3#newcode738
src/ia32/macro-assembler-ia32.cc:738: UpdateAllocationTopHelper(result_end,
scratch);
I assume the order was changed here generate instructions with better scheduling
(it would be nice if the change log said more about why changes were made...) 
The result is probably going to be used almost immediately.  On a previous
compiler I found using lea instead of or_ was faster because the address
generation unit has faster bypass.