d4e3adca57
compiler.cfg.ssa.destruction.coalescing: lookup variable once not twice.
2015-06-15 09:11:16 -07:00
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
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
c0645ecafe
compiler.cfg.linear-scan.*: fix help-lint errors
2015-06-14 16:44:37 +02:00
5ad35fabd1
compiler.cfg.*: lots of more compiler docs
2015-06-14 16:44:37 +02:00
418438ab13
compiler.cfg.linear-scan.*: change compute-live-intervals so that it
...
returns only one sequence, it makes some words a little simpler
2015-06-14 16:44:37 +02:00
f37ab939f0
compiler.*: more compiler docs
2015-06-14 16:44:37 +02:00
91118ae5df
compiler.cfg.linear-scan.allocation.state: fix #1345 by adding vreg to
...
the heap key
By having { start end vreg } as the key in the min-heap ordering is
guaranteed even if two intervals share the same start and end.
2015-06-14 16:44:37 +02:00
9dbb3e9171
compiler.codegen.gc-maps: simplies largest-spill-slot and refactors serialize-gc-maps
2015-06-14 16:44:37 +02:00
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
be342e8638
compiler.cfg.ssa.destruction: refactors maybe/must-eliminate-copy into
...
one try-eliminate-copy
2015-06-14 16:44:36 +02:00
98ae0eb02e
compiler.cfg.ssa.destruction: small refactoring which removes
...
class-elements and maybe-eliminate-copy-later
2015-06-14 16:44:36 +02:00
ceff23339c
compiler.cfg.ssa.destruction: a bunch of new docs
2015-06-14 16:44:36 +02:00
b4f881a9be
compiler.cfg.ssa.destruction.tests: a bunch of new tests
2015-06-14 16:44:36 +02:00
dfa763434e
Fix tests broken by word hashcode change.
2015-06-09 20:37:55 -07:00
d68175f48f
Some more general fixes for recent refactors.
2015-06-08 19:15:29 -07:00
6e60c811ac
core: Rename create to create-word, create-in to create-word-in.
2015-06-08 12:47:35 -07:00
fa2bf33659
compiler.tree: renaming is hard. cleanup -> cleanup-tree
2015-06-05 23:18:43 -07:00
16094d283b
compiler.tree.debugger: more "cleanup"
2015-06-05 23:04:05 -07:00
a0cb7b4fd8
compiler.tree.cleanup: Fix docs
2015-06-05 21:22:50 -07:00
808342076d
compiler: rename cleanup to cleanup-tree, cleanup* to cleanup-tree*
2015-06-05 21:08:18 -07:00
a66bed77b3
compiler.cfg.stacks.clearing: fix help-lint.
2015-05-18 10:02:35 -07:00
c7ea419874
compiler.cfg.stacks.clearing: fix tests.
2015-05-17 09:32:44 -07:00
9ceea3c241
compiler.*: new cfg instruction ##clear, it makes it easier to see where
...
the compiler.cfg.stacks.clearing pass inserts extra instructions
2015-05-17 07:59:34 +02:00
655faadec9
compiler.*: now when we're tracking uninitialized locations instead of
...
initialized ones, the height slot isn't needed on ##call anymore
2015-05-17 07:59:34 +02:00
6dd2fe31b6
compiler.cfg.stacks.*: new attempt at fixing the gc maps bugs
...
the padding vocab is inverted so that instead of trying to track which
locations are live, it tracks those which are dead which is much easier
to get right.
2015-05-17 07:59:34 +02:00
aa1a5f22ba
compiler.tree.debugger: fix docs using.
2015-05-14 08:16:27 -07:00
2c3492a916
minor cleanup to some docs.
2015-05-13 16:09:14 -07:00
296968580a
use assoc-reject instead of [ ... not ] assoc-filter.
2015-05-12 19:08:42 -07:00
b366a06c41
use reject instead of [ ... not ] filter.
2015-05-12 18:50:34 -07:00
f9a6bb27f4
compiler.cfg.*: new test-case and setting leader-map to f is better
2015-05-09 20:10:27 -07:00
409a32d555
compiler.cfg.stacks.padding: docs
2015-05-09 20:04:23 -07:00
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
ecf3829843
compiler.cfg.stacks.*: new vocab 'padding' to perform much more accurate
...
live analysis. it will replace the 'map' vocab
2015-05-09 20:04:22 -07:00
0c2559ef6e
compiler.*: a bunch more compiler docs
2015-05-09 20:04:22 -07:00
e2c91852e4
compiler.cfg.utilities: unused word
2015-05-09 20:04:22 -07:00
8df92abeb9
compiler.*,cpu.*: add a height slot to ##call nodes, then other compiler
...
passes can much easier to accurate liveness analysis on stack locations
2015-05-09 20:04:22 -07:00
8f02cad9c5
compiler.cfg.*: a bunch of new tests
2015-05-09 20:04:22 -07:00
a195e361df
compiler.cfg.stacks.*: create-locs, word for creating stack locations
...
from a sequence
2015-05-09 20:04:21 -07:00
62ea429c2b
compiler.cfg.linear-scan.assignment: Trivial docs fix.
2015-04-30 18:39:31 -07:00
a7cd03b45d
compiler.cfg.linear-scan.assignment: fix help-lint.
2015-04-30 17:20:52 -07:00
5ca23c4793
compiler.*: a few extra doc fixes
2015-04-30 04:51:58 +02:00
6097e80178
compiler.cfg.*: fixing tests so they work on 32bit archs too
2015-04-30 04:49:48 +02:00
b4e7ffede4
compiler.cfg.stacks.local: test and fix for a silly bug in
...
compute-local-kill-set (switched the order on ds and rs)
2015-04-29 09:32:00 -07:00
9fc73c3355
compiler.cfg.stacks.local: unused words
2015-04-29 09:32:00 -07:00
d0c1493f36
compiler.cfg.linearization: can't use slurp/replenish-queue here because
...
the successors need to be pushed to the back of the deque
2015-04-29 09:32:00 -07:00
24ca0e3160
compiler.cfg.*: reverting the commit
...
5fd9e61bc04b054d58b8eeecbe9b5c59f68f2195, caused bad regressions in SIMD code
2015-04-29 09:31:59 -07:00
050e15ce83
compiler.cfg.debugger: new words build-optimized-tree and run-passes to
...
remove some duplicated code
2015-04-29 09:31:59 -07:00
6215561ad4
compiler.cfg.*: more docs and removing a test that isn't relevant anymore
2015-04-29 09:31:59 -07:00
17dff48402
compiler.cfg.ssa.*: refactors words to use stack parameters instead of
...
dynamic variables
2015-04-29 09:31:59 -07:00
2077ae9180
compiler.cfg.linear-scan.assignment: refactoring vreg>reg and
...
assign-registers-in-block so locals aren't needed
2015-04-29 09:31:59 -07:00
135d7f5ff3
compiler.cfg.*: more unit tests
2015-04-29 09:31:59 -07:00
992314d2de
compiler.cfg.*: new word for consuming deques slurp/replenish-deque
...
most uses of slurp-deque processes one item and pushes a sequence of
items to continue working with. it can be formalized into a
slurp/replenish-deque combinator which also reduces the amount of
variables you need to use
2015-04-29 09:31:58 -07:00
c3ca8bd859
compiler.cfg.*: even more compiler docs
2015-04-29 09:31:58 -07:00
997f95196c
compiler.cfg.*: lots of new unit tests
2015-04-29 09:31:58 -07:00
167d1598b3
compiler.cfg.linear-scan.allocation.*: lots more docs and some word simplifications
2015-04-29 09:31:58 -07:00
63a1e90af8
compiler.cfg.ssa.cssa, compiler.cfg.ssa.destruction, compiler.utilities: docs & tests
2015-04-29 09:31:58 -07:00
16f6025979
compiler.cfg.linear-scan.allocation.spilling: docs and tests for this vocab
2015-04-29 09:31:58 -07:00
b827bc75cc
compiler.*: bunch of new docs for some compiler instructions
2015-04-29 09:31:58 -07:00
a17e943e88
compiler.cfg.liveness: refactoring so that words that doesn't modify
...
the live-set doesnt return it
e.g it's clearer when the effect is ( live-set insn -- ) than ( live-set insn -- live-set ) that the live-set is modified in place
2015-04-29 09:31:57 -07:00
c876d79b9d
compiler.cfg.liveness: lots of more tests and some docs
2015-04-29 09:31:57 -07:00
aa6fbf475e
compiler.cfg: the id slot of basic-block appears to not be needed
2015-04-29 09:31:57 -07:00
5cd44dc07c
compiler.cfg.linear-scan.allocation.state: set spill-area-size/align when the cfg tuple is created
2015-04-29 09:31:57 -07:00
d486de889d
compiler.cfg.*: smaller doc improvements
2015-04-29 09:31:57 -07:00
4cda101717
compiler.cfg.linear-scan.allocation.state: more docs and tests for vocab
2015-04-29 09:31:57 -07:00
043b6d5377
compiler.cfg.linear-scan.linear-scan: refactoring of word so that locals doesn't need to be used and more docs (by moving comments)
2015-04-29 09:31:56 -07:00
6c6f85428c
compiler.cfg.intrinsics.simd: refactoring to use assocs instead of case-s for lookups + tests
2015-04-29 09:31:56 -07:00
745d3f29b2
compiler.cfg.*: a bunch of new docstrings and other doc fixes
2015-04-29 09:31:56 -07:00
f9812d9874
compiler.cfg.stacks.*: possible fix for #1289
...
instead of delaying outputting ##replace instructions to when they are
needed, output them directly in the block so that "holes" in the
stacks are avoided.
2015-04-29 09:31:56 -07:00
177a356240
compiler.cfg.stacks.local: refactorings on the vocab to try and make it using less dynamic variables
2015-04-29 09:31:56 -07:00
91144c0712
compiler.cfg.stacks.*: simplify the code a little by making replace-sets, peek-sets and kill-sets contain hash-sets instead of hash-tables
2015-04-29 09:31:56 -07:00
0ff4c68e15
compiler.cfg.stacks.local: to simplify, end-local-analysis can take the block it operates on as argument
2015-04-29 09:31:56 -07:00
b983053130
compiler.cfg.*: putting test-init in compiler.cfg.stacks.tests didn't work, put new word cfg-unit-test in compiler.test
2015-04-29 09:31:55 -07:00
4390561a6e
compiler.cfg.*: new word reset-vreg-counter, useful for making test writing more convenient
2015-04-29 09:31:55 -07:00
2536943e82
compiler.cfg.*: new word connect-Nto1-bbs
...
using that word predecessors is already up-to-date so calling needs-predecessors isn't needed
2015-04-29 09:31:55 -07:00
8a7699e42c
compiler.cfg.*: refactoring to remove the initial-basic-block word
2015-04-29 09:31:55 -07:00
6559382028
compiler.cfg.*: ds-load removed
2015-04-29 09:31:55 -07:00
cc1903bec1
compiler.cfg.*: some refactoring in the stack tracking for #shuffle nodes
...
instead of having separate words for the stacks like inc-d/r and ds/rs-store use generic words that work on either stack
2015-04-29 09:31:55 -07:00
eda9535ce6
compiler.tree.propagation.call-effect: changing back call-effect-unsafe?, changing it made all images to large for some reason
2015-04-29 09:31:55 -07:00
837e098d41
compiler.cfg.*: a little more docs
2015-04-29 09:31:54 -07:00
ba4736ff75
compiler.cfg.stacks.local: change current-height to a two-tuple { { d emit-d } { r emit-r } } it makes the code a bit simpler
2015-04-29 09:31:54 -07:00
c360f0123b
compiler.cfg.stacks: a few unused words removed
2015-04-29 09:31:54 -07:00
09238cee4b
compiler.tree.propagation.call-effect: docs for this vocab
2015-04-29 09:31:54 -07:00
acc1bd3221
compiler.*: removing the check-d and check-r slots from gc-map and adjusting code that depended on those
2015-04-29 09:31:53 -07:00
2cd2aa498a
VM: the compiler doesn't emit any live overinitialized locations anymore
...
Any overinitialized location immediately becomes garbage and is not
reused again. This is very nice because then the gc maps can be
simpler again.
2015-04-29 09:31:53 -07:00
ded3760d46
compiler.*: merge of ##inc-r and ##inc-d into one instruction ##inc with a loc slot
...
this is nice because it leads to simpler code and matches how ##peek and ##replace instructions work
2015-04-29 09:31:53 -07:00
c7d9cb4c3d
compiler.cfg.*: new and improved docs
2015-04-29 09:31:53 -07:00
74ea76a3c9
compiler.cfg: unlikely? slot unused, lets get rid of it
2015-04-29 09:31:53 -07:00
bb504b7c3d
compiler.cfg.dependence: this vocab was only used by scheduling so it isn't needed anymore
2015-04-29 09:31:53 -07:00
b3ebb04d6d
compiler.cfg.*: since the height and scheduling compiler passes aren't used anymore lets remove them
2015-04-29 09:31:53 -07:00
e7fc6a789e
compiler.tree.*: more compiler tree docs
2015-02-05 11:01:05 +00:00
edf1bddd67
compiler.tests.optimizer: Re-enable the minor-gc test case that was
...
failing until bjourne's latest commits.
2015-01-22 13:03:28 -08:00
e4e6c0031a
compiler.cfg.stacks.clearing: Fix help-lint-all.
2015-01-22 12:30:49 -08:00
d2c0ae9fd9
compiler.cfg.stacks.*: more docs
2015-01-22 12:17:21 -08:00
71b054cf39
compiler.cfg.stacks.*: updated tests to match
2015-01-22 12:17:21 -08:00
3c1f223d90
compiler.cfg.stacks.*: use insn# as the key in the stack-record instead of the insn tuple itself
2015-01-22 12:17:21 -08:00
d19869f906
compiler.cfg.stacks.*: many new and fixed tests
2015-01-22 12:17:21 -08:00
d4493858da
compiler.cfg.stacks.*: vacant-peek checking readded
2015-01-22 12:17:20 -08:00
640759ffed
compiler.cfg.*: new step in finalization: clear-uninitialized
...
this step is supposed to analyze the cfg, find where there are peeks
that can cause stack underflow, and insert replace-imm instructions so
that there are no uninitialized stack locations if they do. could fix #1187
2015-01-22 12:17:20 -08:00
fd06e22e9a
compiler.cfg.stacks.map: new vocab for creating an assoc that contains the stack state for each instruction in a cfg
2015-01-22 12:17:20 -08:00
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
ceebc1d6c1
compiler: Fix help-lint for @bjourne's recent awesome docs.
2015-01-17 14:02:59 -08:00
b4e9334d04
compiler.tests.stack-trace: Fix unit test.
2015-01-17 13:57:56 -08:00
c4fbeccfa8
compiler.tests.optimizer: Disable minor-gc test case for now. #1187 .
2015-01-17 11:02:05 -08:00
b6d25ba090
compiler.tests.stack-trace: unit tests against the #1265 bugs
2015-01-15 09:29:14 +00:00
d0dde922bd
compiler.cfg.*: new unit test vocabs
2014-12-31 03:37:17 +01:00
5782855343
compiler.cfg.*: more docs
2014-12-31 03:37:16 +01:00
0432b17409
compiler.cfg.*: refactoring away the compute-global-sets word
2014-12-31 03:37:16 +01:00
79d294e4bf
compiler.cfg.*: docs and more tests
2014-12-31 03:37:16 +01:00
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
e87be7b5c8
compiler.cfg.parallel-copy: docs
2014-12-31 03:37:16 +01:00
24c9dbe160
compiler.cfg.stacks.height: these words are unused
2014-12-31 03:37:16 +01:00
2ccbcd6af6
compiler.cfg.*: more compiler docs
2014-12-31 03:37:16 +01:00
e968c16139
compiler.cfg.stacks.finalize: initial docs
2014-12-31 03:37:16 +01:00
c2c722891e
compiler.cfg.save-contexts: fix help-lint.
2014-12-15 15:11:06 -08:00
2a893a110a
compiler.cfg.linear-scan.allocation.state: don't use slot.syntax.
2014-12-15 13:10:59 -08:00
aee55f1b83
compiler.cfg.linear-scan.assignment: forgot that i removed live-intervals>min-heap
2014-12-15 21:04:53 +01:00
d6c1c075ff
compiler.cfg.linear-scan.allocation: using slurp-heap
2014-12-15 20:48:34 +01:00
84f8f26ce8
compiler.cfg.linear-scan.allocation.state: doc fixing
2014-12-15 20:44:53 +01:00
35d00da765
compiler.cfg.linear-scan.allocation.*: maybe a fix for #1227
...
The register allocator pops from the unhandled-intervals and
unhandled-sync-points heaps. If two live intervals share the same
start slot, then they can be popped in either order leading to
differently generated code. So instead use one
heap (unhandled-min-heap) and both the start and end slot as the heap
key to resolve tie-breakers
2014-12-15 20:44:53 +01:00
173f7c0869
compiler.cfg.scheduling: refactoring to use the simple-optimization combinator
2014-12-15 20:44:07 +01:00
8a011c8366
compiler.cfg.*: more docs
2014-12-15 20:44:07 +01:00
c810677b47
compiler.cfg.linear-scan.*: refactoring to simplify init-allocator and get rid of init-unhandled
2014-12-15 20:44:06 +01:00
fab478482e
compiler.cfg.linear-scan.allocation: refactoring to simplify the (allocate-registers) recursion
2014-12-15 20:44:06 +01:00
52c6009aeb
compiler.cfg.linear-scan.assignment: utility word heap-pop-while which lets you express expire-old-intervals and activate-new-intervals more nicely + tests
2014-12-15 20:40:49 +01:00
0af46ac4de
compiler.cfg.linear-scan.assignment: more docs and refactoring of the init-unhandled and assign-registers-in-block words
2014-12-15 20:40:49 +01:00
a7137f32ff
compiler.cfg.save-contexts: refactoring of the vocab, using the simple-optimization combinator and union classes you can shorten it a lot
2014-12-15 20:36:27 +01:00
e5c69d3e9b
compiler: try this for compiler.cfg.builder.alien.
2014-12-13 18:44:35 -08:00
a689211ae1
compiler.cfg.builder: use compiler.cfg.builder.alien.
2014-12-13 18:08:50 -08:00
c65877104e
compiler: cleanup usings.
2014-12-13 16:10:21 -08:00
46e0426795
compiler.cfg.stack-frame: fix tests.
2014-12-12 17:14:39 -08:00
1911b183a0
compiler.cfg.build-stack-frame: fix tests.
2014-12-12 15:22:22 -08:00
8f4c4ea2e1
compiler.cfg.build-stack-frame: some cleanup.
2014-12-12 11:59:23 -08:00
c3a4bdcf68
compiler.cfg.linear-scan.resolve: use if-empty.
2014-12-11 14:00:50 -08:00
068ddd6c7b
compiler: cleanup cfg passes to have stack effect ( cfg -- )
2014-12-11 12:48:43 -08:00
33500db79f
compiler.cfg.*: help-lint and test case fixes
2014-12-11 15:02:40 +01:00
08e9f18c3b
compiler.cfg.*: fix for test cases that broke due to stack effect changes
2014-12-11 14:30:47 +01:00
bfc0ef815b
compiler.cfg.dominance: fix for #1227 , the dom children needs to be sorted to ensure that the same instruction sequence is generated every time
2014-12-11 14:30:47 +01:00
628c87c5b1
compiler.cfg.utilities: adding connect-bbs and make-edges
2014-12-11 14:30:47 +01:00
00e2d9f116
compiler.cfg.dominance: some more docs
2014-12-11 14:30:47 +01:00
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
e371327ec4
compiler.cfg.ssa.construction.tests: fix tc broken by previous commit
2014-12-11 14:30:47 +01:00
befbe0bb59
compiler.cfg.dominance: changed stack effect compute-dominance ( cfg -- cfg' ) to ( cfg -- )
2014-12-11 14:30:47 +01:00
e5866dfa80
compiler.cfg.*: changed stack effect of needs-predecessors from ( cfg -- cfg') to ( cfg -- )
2014-12-11 14:30:47 +01:00
76cb665a8d
compiler.cfg.dominance: change needs-dominance effect from ( cfg -- cfg') to ( cfg -- )
2014-12-11 14:30:46 +01:00
054982f24a
compiler.cfg.*: docs for compiler.cfg.dominance and compiler.cfg.ssa.construction
2014-12-11 14:30:46 +01:00
53b7f82368
Docs: more compiler docs
2014-12-11 14:30:46 +01:00
5fe9ce4235
compiler.cfg: fix stack effect of cfg-changed and predecessors changed to ( cfg -- )
2014-12-11 14:30:46 +01:00
a0d4e9b417
compiler.cfg.stacks.uninitialized: this compiler pass isn't used anymore -- let's get rid of it
2014-12-11 14:30:46 +01:00
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
John Benediktsson
Björn Lindqvist
Doug Coleman