Joe Groff
|
6d328be69b
|
all-fp-exceptions constant
|
2009-09-12 18:13:25 -05:00 |
Slava Pestov
|
51dd22b119
|
math.floats.env.ppc: fix ppc-fp-traps>bit
|
2009-09-12 12:57:15 -05:00 |
|
Slava Pestov
|
d5206b26ec
|
math.floats.env.ppc: fix ppc-fp-traps-bits
|
2009-09-12 12:49:51 -05:00 |
|
Joe Groff
|
df4d1f8dd2
|
typos in math.floats.env.ppc
|
2009-09-12 09:03:27 -05:00 |
|
Joe Groff
|
61851fcd7a
|
take the union of the x87 and sse exception flags when reporting fp-exception-flags. add back the unit tests i took out since this should fix the problem
|
2009-09-09 23:37:48 -05:00 |
|
Joe Groff
|
3d49cc5a0d
|
change math.floats.env tests not to use any libm functions, which don't reliably set the right hardware exceptions on linux glibc
|
2009-09-09 18:00:38 -05:00 |
|
Joe Groff
|
01b5430fbf
|
make math.floats.env backends unportable
|
2009-09-09 17:32:26 -05:00 |
|
Joe Groff
|
1c62a978d0
|
add some tests to help track down leaks in FP state changes
|
2009-09-08 23:07:33 -05:00 |
|
Joe Groff
|
1fcc0875ba
|
update math.floats.env.x86 for sse detection change
|
2009-09-08 15:47:03 -05:00 |
|
Joe Groff
|
0ec342a1c5
|
factor out x86 and ppc backends for math.floats.env; update both x87 and SSE state on x86
|
2009-09-08 15:36:53 -05:00 |
|
Joe Groff
|
e3509e7f11
|
tighten up math.floats.env docs
|
2009-09-07 10:50:40 -05:00 |
|
Joe Groff
|
25b45d6af0
|
test fp traps
|
2009-09-06 09:04:46 -05:00 |
|
Joe Groff
|
e705470d42
|
make public words for querying current rounding mode, denormal mode, and trap set
|
2009-09-06 08:50:54 -05:00 |
|
Joe Groff
|
96db254311
|
unit tests for math.floats.env
|
2009-09-06 07:50:56 -05:00 |
|
Joe Groff
|
36ec53a968
|
math.floats.env vocab with words to control the floating-point environment
add some functions to the VM to grab and set the fpu control register
|
2009-09-05 19:48:13 -05:00 |