 yury
|
6537b99ac3
* i386: Fixed detection of a peephole optimization using CMOV.
|
9 years ago |
 florian
|
c40240990e
* popt386 uses now also all routines of aoptx86
|
9 years ago |
|
florian
|
53ea4fb7d4
* unify x86 peephole optimizer helpers
|
10 years ago |
|
florian
|
9f587625e7
+ handle not and neg in isFoldableArithOp as well
|
10 years ago |
|
florian
|
de17e9fa1c
+ cpu capability CPUX86_HAS_CMOV
|
10 years ago |
|
florian
|
06fd4223e9
+ optimization setting level4
|
10 years ago |
|
florian
|
35a04ad095
* modified patch of Paul W to replace leave by mov/pop, resolves #26455
|
10 years ago |
 sergei
|
2ee0c8de45
* i386: For integer comparisons with zero, emit "test $-1,%reg" instead of "test %reg,%reg". It is more spilling-friendly, because it transforms into "test $-1,spilltemp" and does not require a register.
|
11 years ago |
|
florian
|
cbefdcf887
* patch by Martin Friebe to avoid wrong debugger output, resolves #25584
|
11 years ago |
 nickysn
|
1de847c754
* fixed error in comment
|
11 years ago |
|
florian
|
4d5119bf1c
* fixes several issues which cause warnings by the dfa code when using it to detect uninitialized variables
|
11 years ago |
 Jonas Maebe
|
0302453912
* disabled the lea->add transformation added in r25907 because it breaks
|
11 years ago |
|
florian
|
1843ed59b5
* set p to a valid instruction
|
11 years ago |
|
florian
|
a16ceb6cdf
* replace lea by add if possible
|
11 years ago |
|
florian
|
80f8f66c6e
* set file positions correctly in PeepHoleOptPass1
|
11 years ago |
|
florian
|
4148637c8c
* fold mov/lea/mov as well if possible
|
11 years ago |
|
florian
|
524cde8140
+ MatchReference
|
11 years ago |
|
florian
|
440cb33e80
* continue to use lea to adjust the stack pointer, it is the recommended way
|
11 years ago |
|
florian
|
5eb578a44c
* generate sub instead of add from lea if appropriate
|
11 years ago |
|
florian
|
d575a435c3
* some debugging infrastructure
|
11 years ago |
|
florian
|
bfbb0c5b9d
* optimize mov/lea
|
11 years ago |
|
nickysn
|
f288859fba
* i386 optimization - remove the test/or %y, %y instruction in the sequence:
|
11 years ago |
|
florian
|
507edb16de
* use add/sub instead of inc/dec on modern CPUs when optimizing for speed
|
12 years ago |
|
florian
|
2ea6979277
* handle Movzl/Op/Mov optimization correctly, resolves #23431
|
12 years ago |
|
florian
|
df01717c38
* mov x, %treg; mov %treg, y might be only optimized if y does not use treg
|
13 years ago |
|
Jonas Maebe
|
dd03bc3e66
* i386 version of fix from r21113
|
13 years ago |
|
florian
|
a6b2fcfe22
* fixes breakage of r17766, should resolve #19598
|
14 years ago |
|
florian
|
b00a610aec
* optimize andb/w const,reg andw/l const,reg sequences as well
|
14 years ago |
|
Jonas Maebe
|
d10f46ec59
* skip instructions containing a reference with a segment prefix (fixes
|
14 years ago |
|
Jonas Maebe
|
8f22a5d224
* perform simple imul->lea transformations also for Pentium 3-class target
|
14 years ago |
