Commit Graph

69 Commits (820207c5b07b1efa7e4db1346794e0ee2fb276a1)

Author SHA1 Message Date
Björn Lindqvist 820207c5b0 compiler.cfg.ssa.destruction.coalescing: simpler code for setting up the
initial leader-map and class-element-map
2015-07-28 17:58:28 -07:00
Björn Lindqvist b4e4792cb6 compiler.cfg.*: a bunch of docs updates 2015-07-28 17:58:27 -07:00
John Benediktsson d4e3adca57 compiler.cfg.ssa.destruction.coalescing: lookup variable once not twice. 2015-06-15 09:11:16 -07:00
Björn Lindqvist aeaed40d9c compiler.cfg.ssa.destruction.coalescing: new vocab to refactor and
simplify the prepare-insn code in compiler.cfg.ssa.destruction
2015-06-15 12:28:31 +02:00
Björn Lindqvist c5ec490f39 compiler.cfg.ssa.destruction: the natural-sort was malplaced, only sort
inputs from ##phi, should fix remainder of #1345
2015-06-15 12:28:30 +02:00
Björn Lindqvist 5ad35fabd1 compiler.cfg.*: lots of more compiler docs 2015-06-14 16:44:37 +02:00
Björn Lindqvist ca2d64af68 compiler.cfg.ssa.destruction: try and sort the vreg pairs before
elimination, this way you always get the same leader info
2015-06-14 16:44:36 +02:00
Björn Lindqvist be342e8638 compiler.cfg.ssa.destruction: refactors maybe/must-eliminate-copy into
one try-eliminate-copy
2015-06-14 16:44:36 +02:00
Björn Lindqvist 98ae0eb02e compiler.cfg.ssa.destruction: small refactoring which removes
class-elements and maybe-eliminate-copy-later
2015-06-14 16:44:36 +02:00
Björn Lindqvist ceff23339c compiler.cfg.ssa.destruction: a bunch of new docs 2015-06-14 16:44:36 +02:00
Björn Lindqvist b4f881a9be compiler.cfg.ssa.destruction.tests: a bunch of new tests 2015-06-14 16:44:36 +02:00
John Benediktsson b366a06c41 use reject instead of [ ... not ] filter. 2015-05-12 18:50:34 -07:00
Björn Lindqvist f9a6bb27f4 compiler.cfg.*: new test-case and setting leader-map to f is better 2015-05-09 20:10:27 -07:00
Björn Lindqvist 24dc0679dc compiler.cfg.ssa.destruction: fix and testcase for the
"lookup-base-pointer* does not define a method for the POSTPONE: f
class" bug
2015-05-09 20:04:22 -07:00
Björn Lindqvist 24ca0e3160 compiler.cfg.*: reverting the commit
5fd9e61bc04b054d58b8eeecbe9b5c59f68f2195, caused bad regressions in SIMD code
2015-04-29 09:31:59 -07:00
Björn Lindqvist 17dff48402 compiler.cfg.ssa.*: refactors words to use stack parameters instead of
dynamic variables
2015-04-29 09:31:59 -07:00
Björn Lindqvist c3ca8bd859 compiler.cfg.*: even more compiler docs 2015-04-29 09:31:58 -07:00
Björn Lindqvist 997f95196c compiler.cfg.*: lots of new unit tests 2015-04-29 09:31:58 -07:00
Björn Lindqvist 63a1e90af8 compiler.cfg.ssa.cssa, compiler.cfg.ssa.destruction, compiler.utilities: docs & tests 2015-04-29 09:31:58 -07:00
Björn Lindqvist d486de889d compiler.cfg.*: smaller doc improvements 2015-04-29 09:31:57 -07:00
Björn Lindqvist 1bd4525ac6 compiler.cfg.stacks.local: refactoring making stack-changes and height-changes take and return stuff instead of using variables and the make building 2014-12-31 03:37:16 +01:00
John Benediktsson c65877104e compiler: cleanup usings. 2014-12-13 16:10:21 -08:00
John Benediktsson 068ddd6c7b compiler: cleanup cfg passes to have stack effect ( cfg -- ) 2014-12-11 12:48:43 -08:00
Björn Lindqvist 76cb665a8d compiler.cfg.dominance: change needs-dominance effect from ( cfg -- cfg') to ( cfg -- ) 2014-12-11 14:30:46 +01:00
Alex Vondrak e90ce0865c compiler.cfg.ssa.destruction.leaders: break off leader-map to resolve circular dependencies (oops) 2012-12-31 11:20:57 -08:00
Alex Vondrak fc91dcf530 compiler.cfg.ssa.destruction: alien-call-insns are too hairy to coalesce (issue #22) 2012-12-31 11:04:57 -08:00
Slava Pestov 07227f22f9 WIP 2012-12-31 11:03:33 -08:00
Slava Pestov ec2f42fd40 compiler.cfg.liveness: merge in compiler.cfg.liveness.ssa and simplify the code, since we don't compute live sets before SSA construction anymore 2010-09-25 14:36:58 -07:00
Slava Pestov 355d89e8e8 compiler.cfg: now that kill-blocks cannot contain instructions that define vregs we can skip them all 2010-07-27 12:40:31 -04:00
Slava Pestov e27adb2830 compiler: re-architect low-level optimizer to allow more than one output value per instruction 2010-07-13 07:40:14 -04:00
Slava Pestov 41d929a201 compiler.cfg.ssa.interference: implement linear-time interference test 2010-05-17 05:50:13 -04:00
Slava Pestov 5ecd2b61c7 compiler.cfg: more silly optimizations 2010-05-14 18:37:09 -04:00
Slava Pestov e90712b1ed compiler.cfg: cleanup 2010-05-03 17:34:26 -04:00
Slava Pestov 65f86bfce7 compiler.cfg.ssa.detruction: coalesce different representations in more cases 2010-05-03 17:34:22 -04:00
Slava Pestov 43f269e4eb Register allocation now uses SSA properties to coalesce values with different representations 2010-05-03 17:34:20 -04:00
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
Slava Pestov 82fb1879af Debugging untagged fixnums 2010-05-03 17:34:02 -04:00
Daniel Ehrenberg 7074979745 Moving new-sets to sets 2010-02-26 16:01:01 -05:00
Slava Pestov 299b5b0f6c filter-here -> filter! 2009-10-28 00:44:05 -05:00
Slava Pestov 810cd7b4bb compiler.cfg.ssa.destruction: don't coalesce registers with different representations, even if they're in the same register class, since this messes up spilling 2009-09-29 22:28:20 -05:00
Slava Pestov 08a2eb74f4 cpu.x86: shifts didn't work if dst != src1; re-organize file a bit 2009-09-28 05:39:53 -05:00
Slava Pestov da5f0d0a14 compiler.cfg.ssa.destruction: we can coalesce two registers with different representations as long as they both belong to the same register class 2009-09-28 03:24:54 -05:00
Slava Pestov 1e841e5086 compiler.cfg.ssa.destruction: more aggressive coalescing work in progress 2009-09-27 17:17:26 -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 4b7ba38aab compiler.cfg: virtual registers are integers now, and representations are stored off to the side. Fix bug in representation selection that would manifest if a value was used as a float and a fixnum in different branches; cannot globally unbox float in this case 2009-08-08 04:02:18 -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 dac7edd2ab compiler.cfg.def-use: remove compute-def-use word, passes have to call compute-defs or compute-uses now; compiler.cfg.ssa.liveness: don't compute dominance and def-use first since destruction does already 2009-08-02 19:12:32 -05:00