Joel Dice
b4169ede91
fix reversed arguments to maybeRex in sseMoveRR
...
This was causing 8-byte SSE-to-SSE moves involving registers
xmm8-xmm15 to be misencoded on x86_64, leading to incorrect code
generation in methods with lots of local variables of type double.
2011-07-06 11:09:06 -06:00
Joel Dice
04454960ec
various bugfixes for handling 64-bit floating point values on 32-bit systems
2009-11-30 15:10:34 +00:00
Joel Dice
5ead8fab17
refactor code responsible for moving data in the compiler
...
This is partially to address incorrect code generation for 64-bit
floating-point values on x86_32 and partially to reduce unnecessary
moves.
2009-11-27 21:15:12 -07:00
Joel Dice
fb5796b740
don't use SSE for long-to-double conversion on 32-bit systems
2009-11-04 00:02:38 +00:00
Joel Dice
064ad572bc
use truncation, not rounding in float2Int instructions
2009-10-20 15:43:24 -06:00
Joel Dice
4f78783ef1
various bugfixes for SSE-based floating-point support
2009-10-05 14:25:12 +00:00
Joel Dice
ed9e54bca1
add a few more tests to Floats.java
2009-03-04 17:52:06 -07:00
Joel Dice
f05318b787
fix thinkos in d2i and friends
2008-05-22 18:08:41 -06:00
Joel Dice
5f6258045e
implement various floating point instructions; fix Floats.java to actually test something
2007-10-15 13:12:38 -06:00
Joel Dice
89609e11c9
implement floating point instructions
2007-08-13 08:06:31 -06:00
