026abb9c38
compiler.cfg.linear.scan.*: makes it so ranges and uses are guaranteed
...
to be vectors
Simplifies the code and makes it a little faster.
2016-04-02 17:48:23 +02:00
76be7b54ed
compiler.cfg.linear-scan.*: removed start and end from live-interval-state
...
They weren't needed because start>> was always equal to ranges min and
end>> ranges max
2016-04-01 19:16:35 +02:00
7b18618b3e
compiler.cfg.linear-scan.allocation.state: fixes the the 64-bit-o-centric testcases
2015-11-04 13:54:14 +01:00
f692200547
compiler.cfg.*: refactoring which removes spill-area-size and
...
spill-area-align from cfg
This makes the code for building the stack frame simpler.
2015-11-04 07:53:19 +01:00
cfe4fec574
compiler.cfg.linear-scan.*: make it so cfg is given as an argument to next-spill-slot
2015-11-04 07:53:19 +01:00
fb45bc278f
compiler.cfg.linear-scan.live-intervals: the reg-class slot can be
...
removed
it's always the same as vreg>> rep-of reg-class-of so no need to store
that data twice
2015-09-22 08:51:04 +02:00
22b871d82e
compiler.cfg.utilities: make it so block>cfg initializes spill-area-align
2015-07-29 13:01:24 +02:00
b544c1b352
compiler.cfg.*: some extra unit tests for the compiler
2015-07-28 17:58:29 -07:00
40892c0302
tests: prefer ${ } instead of 1array
2015-07-02 14:56:14 -07:00
59f3b1ea57
Revert "factor: rename [ ] [ ] unit-test -> { } [ ] unit-test using a refactoring tool!"
...
Needs a bit more work.
This reverts commit 7e54cc2824
2015-07-02 13:47:06 -07:00
7e54cc2824
factor: rename [ ] [ ] unit-test -> { } [ ] unit-test using a refactoring tool!
2015-07-02 11:36:08 -07:00
8c8605d11a
compiler.cfg.linear-scan.*: more tests and docs for allocation and state
2015-06-16 18:28:32 -07:00
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
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
6097e80178
compiler.cfg.*: fixing tests so they work on 32bit archs too
2015-04-30 04:49:48 +02:00
997f95196c
compiler.cfg.*: lots of new unit tests
2015-04-29 09:31:58 -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
4cda101717
compiler.cfg.linear-scan.allocation.state: more docs and tests for vocab
2015-04-29 09:31:57 -07:00
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
1eac4afc8b
compiler.cfg.linear-scan.allocation.state: sub docs and tests
2014-11-16 06:25:53 -08:00
Björn Lindqvist
Doug Coleman