Björn Lindqvist
|
163fef12f9
|
compiler.cfg.*: detached the normalize-height and schedule-instructions passes
those stages cause uninitialized "holes" in the stacks which in turn
causes gc errors if stack underflows occur
|
2015-01-22 12:17:20 -08:00 |
John Benediktsson
|
068ddd6c7b
|
compiler: cleanup cfg passes to have stack effect ( cfg -- )
|
2014-12-11 12:48:43 -08:00 |
Björn Lindqvist
|
c85175c10c
|
compiler.cfg.*: new utility word apply-passes to improve the looks of long cleave's where each quotation only contains one word
|
2014-12-11 14:30:47 +01:00 |
Björn Lindqvist
|
8744d908b6
|
compiler.cfg.*: fix stack effect of compiler passes in optimize-cfg
since they all modify the cfg in place, it is better if their stack
effect is ( cfg -- ) instead of ( cfg -- cfg' )
|
2014-12-11 14:30:45 +01:00 |
Slava Pestov
|
c677c35de4
|
compiler.cfg: fix major facepalm with write barrier elimination
|
2010-09-05 21:39:45 -07:00 |
Daniel Ehrenberg
|
277705345f
|
Merge branch 'master' of git://factorcode.org/git/factor into s3
Conflicts:
basis/compiler/cfg/finalization/finalization.factor
|
2010-05-03 17:29:03 -05:00 |
Slava Pestov
|
4352902bb6
|
compiler.cfg.empty-blocks: remove unused pass
|
2010-05-03 18:23:54 -04:00 |
Daniel Ehrenberg
|
5509604ffe
|
Merge branch 'master' into s3
Conflicts:
basis/compiler/cfg/optimizer/optimizer.factor
|
2010-05-03 17:19:28 -05:00 |
Slava Pestov
|
9b34a4a054
|
compiler.cfg: remove unused 'reps' slot from compiler.cfg, and re-organize things in preparation for SSA register allocation
|
2010-05-03 17:34:18 -04:00 |
Slava Pestov
|
95ff5ffe51
|
New GC checks work in progress
|
2010-05-03 17:34:16 -04:00 |
Daniel Ehrenberg
|
e22f66daf4
|
Scheduling has to come before representation selection, since the latter inserts memory operations
|
2010-04-05 00:24:12 -05:00 |
Daniel Ehrenberg
|
2034e1ef05
|
Fewer unnecessary dependence edges between stack operations, implemented by reviving stack height normalization. Now, 40% of spills and reloads are eliminated in total
|
2010-02-24 14:20:43 -06:00 |
Daniel Ehrenberg
|
04946dcddd
|
A first attempt at implementing Vivek Sarkar's algorithm for scheduling for register pressure
|
2010-02-10 21:43:30 -06:00 |
Slava Pestov
|
1e841e5086
|
compiler.cfg.ssa.destruction: more aggressive coalescing work in progress
|
2009-09-27 17:17:26 -05:00 |
Slava Pestov
|
daff63071f
|
Fix the build
|
2009-09-09 13:44:54 -05:00 |
Slava Pestov
|
7e2f0e5dbf
|
compiler: separate ##save-context instruction from ##alien-invoke, generate a ##save-context for libm calls, and add a pass to combine multiple context saves within a basic block. Fixes crashes with FP traps thrown by libm functions on x86-32
|
2009-09-08 21:50:55 -05:00 |
Doug Coleman
|
3f3d57032b
|
Delete empty unit tests files, remove 1- and 1+, reorder IN: lines in a lot of places, minor refactoring
|
2009-08-13 19:21:44 -05:00 |
Slava Pestov
|
d0c393aa60
|
compiler.cfg: new system to track when results of analyses need to be recomputed (reverse post order, linear order, predecessors, dominance, loops). Passes can now call needs-predecessors, needs-dominance, needs-loops at the beginning, and cfg-changed, predecessors-changd at the end. Linearization order now takes loop nesting into account, and linear scan now uses linearization order instead of RPO.
|
2009-08-08 20:02:56 -05:00 |
Slava Pestov
|
e21ca289c3
|
compiler.cfg.representations: new pass to make global unboxing decisions, relies on new compiler.cfg.loop-detection pass for loop nesting information
|
2009-08-08 00:24:46 -05:00 |
Slava Pestov
|
725280d424
|
Split off the notion of a register representation from a register class
|
2009-08-07 17:44:50 -05:00 |
Slava Pestov
|
370f4c081d
|
compiler.cfg: convert code into two-operand form before SSA destruction; SSA destruction now operates on a relaxed SSA form where multiple defs of the same vreg are allowed, but only within a single basic block. This makes linear scan's coalescing redundant, allowing it to be removed completely
|
2009-08-05 18:57:46 -05:00 |
Slava Pestov
|
9afa39aa3a
|
compiler.cfg.ssa.destruction: rename coalesce word to destruct-ssa
|
2009-07-28 12:56:33 -05:00 |
Slava Pestov
|
d10993b837
|
compiler.cfg: Rename ssa to ssa.construction, coalescing to ssa.destruction
|
2009-07-28 09:34:08 -05:00 |
Slava Pestov
|
d394dd3c36
|
compiler.cfg.branch-splitting: re-enable with a better heuristic
|
2009-07-28 08:48:05 -05:00 |
Slava Pestov
|
c5d7ed58a5
|
Debugging compiler.cfg.coalescing
|
2009-07-27 19:24:13 -05:00 |
Slava Pestov
|
ba696b68b8
|
compiler.cfg.coalescing: more or less complete, now needs debugging
|
2009-07-27 02:20:45 -05:00 |
Slava Pestov
|
7068de6cd3
|
compiler.cfg.copy-prop: Global copy propagation
|
2009-07-24 05:30:46 -05:00 |
Slava Pestov
|
d947c61bd7
|
compiler.cfg.stacks: now performs online local DCN
|
2009-07-23 20:54:38 -05:00 |
Slava Pestov
|
747a2d72c8
|
compiler.cfg.empty-blocks: new pass to delete empty blocks, runs after phi elimination
|
2009-07-23 18:02:46 -05:00 |
Slava Pestov
|
44bcd258f6
|
Insert _loop-entry in linearization pass instead of in CFG builder, so that optimizations don't have to worry about it
|
2009-07-22 06:05:17 -05:00 |
Slava Pestov
|
25c42d04fd
|
compiler.cfg: Fix conflicts after merging local optimization refactoring into DCN branch
|
2009-07-22 03:17:53 -05:00 |
Slava Pestov
|
21a012e3d7
|
compiler.cfg: Major restructuring -- do not compute liveness before local optimization, and instead change local optimizations to be more permissive of undefined values. Now, liveness is only computed once, after phi elimination and before register allocation. This means liveness analysis does not need to take phi nodes into account and can now use the new compiler.cfg.dataflow-analysis framework
|
2009-07-22 03:08:28 -05:00 |
Slava Pestov
|
b869e1250c
|
compiler.cfg.stack-analysis: Out with the old, in with the new
|
2009-07-22 00:52:25 -05:00 |
Slava Pestov
|
e49de006c4
|
compiler.cfg.dcn: Fixing various bugs, hook up with optimizer
|
2009-07-21 22:24:50 -05:00 |
Slava Pestov
|
343151fd62
|
compiler.tree.finalization: undo a change that resulted in a compile time regression
|
2009-07-18 03:33:24 -05:00 |
Slava Pestov
|
5061db2a48
|
compiler.cfg: Move tail call optimization to the end, compiler.tree: simplify finalization pass because of this
|
2009-07-17 22:57:28 -05:00 |
Slava Pestov
|
9f926ab88c
|
compiler.cfg.block-joining: join basic blocks connected by a single edge to improve effectiveness of local optimizations
|
2009-07-16 02:17:58 -05:00 |
Slava Pestov
|
11731f8d48
|
compiler.cfg.value-numbering: merge in compiler.cfg.branch-folding
|
2009-07-14 20:05:01 -05:00 |
Slava Pestov
|
d7aeae45be
|
compiler.cfg.branch-splitting: split blocks with successors
|
2009-07-13 10:44:08 -05:00 |
Slava Pestov
|
608fb054f2
|
compiler.cfg: Some code cleanups, update stack-analysis and phi-insertion to work on CFGs with critical edges
|
2009-07-12 22:22:46 -05:00 |
U-SLAVA-DFB8FF805\Slava
|
508b7272b5
|
compiler.cfg.optimizer: fix irrelevant test
|
2009-07-09 06:17:10 -05:00 |
Doug Coleman
|
f11ba00667
|
Revert "fix a unit test so it's sure to overflow 64bit machines"
This reverts commit 88a545a215 .
|
2009-07-02 18:18:50 -05:00 |
Doug Coleman
|
559a77be43
|
fix a unit test so it's sure to overflow 64bit machines
|
2009-07-02 18:05:55 -05:00 |
Doug Coleman
|
e11e3cf6d2
|
unit test to make sure adds are not optimized incorrectly
|
2009-07-02 18:01:00 -05:00 |
Slava Pestov
|
2de41f01da
|
compiler.cfg.optimizer: enable branch folding
|
2009-07-02 01:05:59 -05:00 |
Slava Pestov
|
f894baf60e
|
compiler.cfg.optimizer: disable branch folding for now
|
2009-06-30 23:07:27 -05:00 |
Slava Pestov
|
3385e50c43
|
compiler.cfg.branch-folding: fold away branches where both inputs are the same register. This results in 'push' being compiled slightly better
|
2009-06-30 22:43:32 -05:00 |
Slava Pestov
|
f412bbc6d6
|
Change branch splitting around a bit, rename useless-blocks to useless-conditionals and rewrite it
|
2009-06-30 22:11:15 -05:00 |
Slava Pestov
|
1acf243cce
|
Fix conflict
|
2009-06-30 20:16:09 -05:00 |
Slava Pestov
|
5c6c3ecd85
|
compiler.cfg.tco: Tail call optimization moved out of compiler.cfg.builder into its own pass
|
2009-06-30 20:13:35 -05:00 |