Improve bit tests on IA32.

Improve bit tests on IA32.
Bit tests with a memory operand should be avoided on Intel processors.
On Atom bts with a memory operand has 10~11 cycles latency and on Core 2
it has 5~6 cycles latency, and also produces multiple uops.
If the bit base operand is a register the latency is 1 cycle instead.
This CL improves several cases from Kraken 1.1 benchmark on Atom. For example,
12% for imaging-desaturate, 8% for audio-fft, and 5% for audio-beat-detection.

Committed: http://code.google.com/p/v8/source/detail?r=7987

[Søren Thygesen Gjesse, 2011-05-23 08:26:17]

LGTM with additional comment.

http://codereview.chromium.org/7048003/diff/1/src/ia32/macro-assembler-ia32.cc
File src/ia32/macro-assembler-ia32.cc (right):

http://codereview.chromium.org/7048003/diff/1/src/ia32/macro-assembler-ia32.c...
src/ia32/macro-assembler-ia32.cc:75: // Set dirty mark for region.
Please add a comment here why loading to a register.

[yuqiang.xian%intel.com, 2011-05-23 08:48:49]

Comment addressed. Thanks for the review.

[Søren Thygesen Gjesse, 2011-05-23 10:47:20]

LGTM

I will commit it.