Add roundsd instruction to ia32 and use it in optimized Math.floor.

Add roundsd instruction to ia32 and use it in optimized Math.floor.

It is available platforms that have SSE 4.1 and allows us to handle
negative numbers without deoptimization. Before we would deoptimize
on negative inputs to Math.floor. x64 already uses this instruction.

* Change Math.floor unit test to make sure every test case gets
optimized by changing the source code for each test case.

* Fix HIR debug printing for some instructions.
Committed: http://code.google.com/p/v8/source/detail?r=8921

[Sven Panne, 2011-08-12 09:41:40]

LGTM if the -0 handling is really needed (a short comment why would be good
then).

General remark: The spec states "The value of Math.ceil(x) is the same as the
value of -Math.floor(-x).", so can we do something similar there?

http://codereview.chromium.org/7628017/diff/4002/src/ia32/full-codegen-ia32.cc
File src/ia32/full-codegen-ia32.cc (left):

http://codereview.chromium.org/7628017/diff/4002/src/ia32/full-codegen-ia32.c...
src/ia32/full-codegen-ia32.cc:3833: // TODO(svenpanne): Allowing format strings
in Comment would be nice here...
Not really related to the current patch, but I guess we can remove this TODO,
anyway... ;-)

http://codereview.chromium.org/7628017/diff/4002/src/ia32/lithium-codegen-ia3...
File src/ia32/lithium-codegen-ia32.cc (right):

http://codereview.chromium.org/7628017/diff/4002/src/ia32/lithium-codegen-ia3...
src/ia32/lithium-codegen-ia32.cc:2720: // Deoptimize on negative zero.
Why do we have to deoptimize on -0 when using roundsd?