Justin Ethier
38847f11cf
Add extra condition to (alloca?)
...
Make this match what is actually done in cgen for when a while statement is introduced, there was an older case not being handled.
2019-03-27 13:18:31 -04:00
Justin Ethier
27469be2ef
Fix typo
2019-03-27 13:18:17 -04:00
Justin Ethier
5c28ba7730
Avoid using alloca where possible for clo's
2019-03-22 13:09:42 -04:00
Justin Ethier
e68fa56f94
Bugfix - account for the use-alloca flag
2019-03-22 12:26:10 -04:00
Justin Ethier
f13186ee17
Use static alloc for clo/vec, instead of alloca
...
This improves performance as the C compiler can better optimize the generated code.
2019-03-22 10:34:48 -04:00
Justin Ethier
b73f68cc8a
Qualify static function
2019-03-21 22:59:58 -04:00
Justin Ethier
3706647583
Refactoring
2019-03-21 17:16:21 -04:00
Justin Ethier
17cecfe834
WIP
2019-03-21 17:08:16 -04:00
Justin Ethier
a8a1fd211d
Temp test file
2019-03-21 13:11:48 -04:00
Justin Ethier
c357663f0a
Issue #310 - Fix type checking for member/assoc
2019-03-20 22:56:33 -04:00
Justin Ethier
12d6cbce5a
Use explicit if's instead of memv for case
...
Theory is the generated code can make better use of the processor cached when using explicit if statements than aggregating over a list via memv, since pairs in the list can be located anywhere across memory.
2019-03-20 17:39:55 -04:00
Justin Ethier
483cdd2758
Issue #309 - Link to local copy of the MTA paper
2019-03-19 12:49:15 -04:00
Justin Ethier
2704ac1ed2
Update README.md
2019-03-19 12:47:28 -04:00
Justin Ethier
f7acb6260d
Add files via upload
2019-03-19 12:42:32 -04:00
Justin Ethier
5dd5a5eef2
Stub for new version
2019-03-17 18:48:27 -04:00
Justin Ethier
1967ce7a93
Added a section for 0.10
2019-03-17 18:47:54 -04:00
Justin Ethier
41eaf8008b
Add "bin" support to Cyc-installation-dir
2019-03-17 18:45:55 -04:00
Justin Ethier
1f7212a964
Faster version of _list-index
2019-03-14 15:43:36 -04:00
Justin Ethier
a72e91b85c
New releas
2019-03-04 17:48:39 -05:00
Justin Ethier
10342c1d2e
Allow 'bench' directive to return success value
2019-03-04 13:32:34 -05:00
Justin Ethier
d964b037a7
Revised previous comments
2019-02-28 15:59:40 -05:00
Justin Ethier
99d5614339
Faster version of vector
2019-02-28 14:25:03 -05:00
Justin Ethier
3bf2a6125c
Fast list/vector functions also allocate mutable objs
2019-02-28 13:47:32 -05:00
Justin Ethier
f64d429ab8
WIP
2019-02-28 09:43:58 -05:00
Justin Ethier
d9f63a85b4
Revert previous change
2019-02-27 16:03:19 -05:00
Justin Ethier
1f8a824bda
More efficient record type construction
2019-02-27 15:44:50 -05:00
Justin Ethier
8772c5355b
Faster initialization of objects create via define-record-type
2019-02-27 14:31:14 -05:00
Justin Ethier
0f0beb024c
Do not use vector-set when creating record types
...
This improves performance as mutations force the GC to do more work.
2019-02-27 10:29:18 -05:00
Justin Ethier
f5b68ee8d2
Handle hashing of complex nums with a 0i component
2019-02-19 13:20:29 -05:00
Justin Ethier
6bfff4b6ed
Perform fixnum type check
2019-02-19 13:17:33 -05:00
Justin Ethier
db7fd2bf81
Tentative version bump
2019-02-19 13:14:05 -05:00
Justin Ethier
ed2ab18be1
Cleanup 0.9.9 fixes
2019-02-18 22:38:52 -05:00
Justin Ethier
8efbf60528
Must explicitly call (complex?) now
2019-02-18 22:12:55 -05:00
Justin Ethier
ec7c3c3308
Revert change to (integer?) for now
...
Causes unit test failures, need to sort this out separately.
2019-02-18 17:18:02 -05:00
Justin Ethier
b626c4c6b7
Removed extraneous char
2019-02-18 16:51:43 -05:00
Justin Ethier
990cc96e4b
Adjust formatting
2019-02-18 16:50:58 -05:00
Justin Ethier
c1af2cfdcc
Issue #303
2019-02-18 16:49:45 -05:00
Justin Ethier
204e31e362
Issue #303 - Handle numeric predicate edge cases
...
Add edge cases to (real?) for complex numbers, and (integer?) for reals.
2019-02-18 16:48:11 -05:00
Justin Ethier
5f0fa5eeea
Revised 0.9.9 section
2019-02-18 13:09:41 -05:00
Justin Ethier
b4c19b1c7d
Use 0.9.9 instead
2019-02-18 12:32:57 -05:00
Justin Ethier
b2997de54a
Prep 0.9.8.1 bugfix release
2019-02-18 12:13:03 -05:00
Justin Ethier
102244be21
Issue #304 - gc_copy_obj must to populate bignums
...
Even without bignums in the nursery we still need this code in place since gc_alloc calls it, and otherwise all of the callers would need to ensure bignums are properly initialized. There may be an opportunity here for optimization, but let's make sure everything works first!
2019-02-18 12:09:31 -05:00
Justin Ethier
b2b10ca729
WIP - new version of libtommath
2019-02-17 19:00:47 -05:00
Justin Ethier
2e9a5bd9c2
Releasing 0.9.8
2019-02-15 18:04:06 -05:00
Justin Ethier
0a0f51ada5
Relocated test files
2019-02-15 17:56:26 -05:00
Justin Ethier
fa01202df8
Do not attempt to memoize functions that take 0 args
2019-02-15 12:36:54 -05:00
Justin Ethier
7021855eca
Inject additional imports, added a TODO
2019-02-14 23:03:33 -05:00
Justin Ethier
d0d7fae89c
Only memoize pure functions for programs
...
The problem with doing this for libraries is that it adds a new dependency on (srfi 69). It would be problematic to use a library with these changes as there is no easy way to know this SRFI is also a library dependency.
2019-02-14 22:53:12 -05:00
Justin Ethier
ae768b0e83
Added feature section
2019-02-14 21:11:54 -05:00
Justin Ethier
e5be007539
Added memoization optimization flags
2019-02-14 21:10:06 -05:00