 Dan Gohman
|
54e4c36a73
Rewrite the SDep class, and simplify some of the related code.
|
16 years ago |
|
Dan Gohman
|
2ce7f2068f
Drop the reg argument to isRegReDefinedByTwoAddr, which was redundant.
|
16 years ago |
|
Dan Gohman
|
490b1833a9
Ignore IMPLICIT_DEF instructions when computing physreg liveness.
|
16 years ago |
|
Dan Gohman
|
80e201b8db
Use register names instead of numbers in debug output.
|
16 years ago |
|
Dan Gohman
|
cef874ae2b
Rewrite the liveness bookkeeping code to fix a bunch of
|
16 years ago |
|
Dan Gohman
|
e96cc775e4
Fix an inconsistency in a comment.
|
16 years ago |
|
Dan Gohman
|
e561751e98
Don't charge the full latency for anti and output dependencies. This is
|
16 years ago |
|
Dan Gohman
|
0dba0e5408
When looking for anti-dependences on the critical path, don't bother
|
16 years ago |
|
Dan Gohman
|
ebb0a31103
Add a comment about callee-saved registers.
|
16 years ago |
 Nick Lewycky
|
a89d102b32
Silence a warning.
|
16 years ago |
|
Dan Gohman
|
878ef1d5bc
Suppress warnings.
|
16 years ago |
|
Dan Gohman
|
21d9003087
Initial support for anti-dependence breaking. Currently this code does not
|
16 years ago |
|
Dan Gohman
|
a1e6d363e5
Factor out the code for verifying the work of the scheduler,
|
17 years ago |
|
Dan Gohman
|
343f0c0467
Experimental post-pass scheduling support. Post-pass scheduling
|
17 years ago |
|
Dan Gohman
|
ae73dc1448
Tidy up several unbeseeming casts from pointer to intptr_t.
|
17 years ago |
 Chris Lattner
|
459525df1e
don't create the post-ra scheduler unless it is enabled.
|
17 years ago |
|
Chris Lattner
|
4ee451de36
Remove attribution from file headers, per discussion on llvmdev.
|
17 years ago |
 Dale Johannesen
|
72f1596403
Modify previous patch per review comments.
|
18 years ago |
|
Dale Johannesen
|
e7e7d0d7e3
Skeleton of post-RA scheduler; doesn't do anything yet.
|
18 years ago |
