Doug Coleman
|
330c7ba156
|
factor: Removing SLOT: nth/at/global in favor of using change: at for tuple slots named ``at`` and change-at for hashtables.
Redo many FROM: and QUALIFIED: and using lists.
|
2015-08-12 08:49:02 -05:00 |
Alex Vondrak
|
d58d467434
|
compiler.cfg.gvn: experiment to scan for available vregs when rewriting
instructions instead of relying on the availability of the canonical
leader; untested
|
2012-09-19 17:28:13 -07:00 |
|
Alex Vondrak
|
fb83bf2d6a
|
compiler.cfg.gvn.avail: add a FROM: to clarify collision over change-at word
|
2012-09-12 15:14:12 -07:00 |
|
Alex Vondrak
|
2f1fdaad6c
|
compiler.cfg.gvn: forgot to make certain defs available
|
2012-09-12 15:14:11 -07:00 |
|
Alex Vondrak
|
b99a90f138
|
compiler.cfg.gvn.avail: make with-available-uses? short-circuit
|
2012-09-12 15:14:10 -07:00 |
|
Alex Vondrak
|
a9aebc2854
|
compiler.cfg.gvn.avail: kludgey fix for odd problem with updating avail sets
|
2012-09-12 15:14:10 -07:00 |
|
Alex Vondrak
|
b90187a7b0
|
compiler.cfg.gvn: try to clean up availability issues (untested)
|
2012-09-12 15:14:10 -07:00 |
|
Alex Vondrak
|
25b0fb5c03
|
compiler.cfg.gvn.avail: minor simplification
|
2012-09-12 15:14:08 -07:00 |
|
Alex Vondrak
|
a3ea32ed3e
|
compiler.cfg.gvn.redundancy-elimination: horrific tinkering that doesn't even work
|
2012-09-12 15:14:08 -07:00 |