]>
gitweb.factorcode.org Git - factor.git/log
Daniel Ehrenberg [Fri, 14 May 2010 23:59:39 +0000 (18:59 -0500)]
Merge branch 'master' of git://factorcode.org/git/factor into propagation
Conflicts:
basis/compiler/tree/propagation/propagation-tests.factor
Joe Groff [Fri, 14 May 2010 23:54:06 +0000 (16:54 -0700)]
gpu.buffers: "with-mapped-buffer-array" helper word that wraps a mapped GL buffer in a specialized-array for easy inspection
Joe Groff [Fri, 14 May 2010 23:17:03 +0000 (16:17 -0700)]
cuda.gl: add free-resource word to release interop handles
Joe Groff [Fri, 14 May 2010 22:59:31 +0000 (15:59 -0700)]
cuda: add "cuda.gl" vocab with words for cuda/opengl/gpu interop
Joe Groff [Fri, 14 May 2010 22:17:26 +0000 (15:17 -0700)]
cuda: in with-cuda-context, sync context when cleaning up so that destroying context doesn't fail due to asynchronous errors
Joe Groff [Fri, 14 May 2010 21:29:48 +0000 (14:29 -0700)]
gpu.render: allow uchar-array, ushort-array, and uint-array to be passed directly to render as element arrays
Slava Pestov [Fri, 14 May 2010 22:31:12 +0000 (18:31 -0400)]
compiler: fix bad unit test
Slava Pestov [Fri, 14 May 2010 22:18:39 +0000 (18:18 -0400)]
cpu.x86.bootstrap: use TEST instruction
Slava Pestov [Fri, 14 May 2010 22:18:29 +0000 (18:18 -0400)]
compiler.cfg: more silly optimizations
Slava Pestov [Fri, 14 May 2010 22:18:04 +0000 (18:18 -0400)]
compiler.cfg.instructions: change vreg-insn from a mixin into a superclass
Slava Pestov [Fri, 14 May 2010 21:55:00 +0000 (17:55 -0400)]
compiler.cfg.linear-scan: pointless optimizations
Slava Pestov [Fri, 14 May 2010 20:44:03 +0000 (16:44 -0400)]
compiler.cfg.ssa.live-ranges: clean up
Slava Pestov [Fri, 14 May 2010 20:43:14 +0000 (16:43 -0400)]
bootstrap.compiler.timing: small update
Slava Pestov [Fri, 14 May 2010 13:08:50 +0000 (09:08 -0400)]
compiler.cfg.alias-analysis: simplify and speed up
Slava Pestov [Fri, 14 May 2010 12:14:22 +0000 (08:14 -0400)]
compiler.cfg: use x86 TEST instruction to optimize 'bitand 0 ='
Joe Groff [Fri, 14 May 2010 20:29:37 +0000 (13:29 -0700)]
sequences.cords: specialize vshuffle-elements to use vshuffle2-elements on cord components
Joe Groff [Fri, 14 May 2010 09:47:39 +0000 (02:47 -0700)]
math.vectors.simd.intrinsics: fix scalar fallback for (simd-vshuffle2-elements)
Joe Groff [Fri, 14 May 2010 09:47:05 +0000 (02:47 -0700)]
math.vectors, math.vectors.simd: add user-facing vshuffle2 word
Joe Groff [Fri, 14 May 2010 08:20:05 +0000 (01:20 -0700)]
compiler.cfg.intrinsics.simd: mod shuffle indices for shuffle-2-vectors-imm intrinsic so they wrap like a real instruction would
Joe Groff [Fri, 14 May 2010 08:16:29 +0000 (01:16 -0700)]
math.vectors.simd.intrinsics: (simd-vshuffle2-elements) intrinsic that creates a vector by selecting elements from two input vectors. use ##shuffle-vector-halves-imm to implement for double-2s with SSE
Joe Groff [Fri, 14 May 2010 07:15:29 +0000 (00:15 -0700)]
compiler.cfg.instructions: ##shuffle-vector-halves-imm insn to map to SSE's two-input SHUFPS/SHUFPD
Slava Pestov [Fri, 14 May 2010 01:55:19 +0000 (21:55 -0400)]
math.vectors.simd.cords: implement new-sequence and like methods on cords to make cross product work
Slava Pestov [Thu, 13 May 2010 22:25:46 +0000 (18:25 -0400)]
cpu.x86.32: check for SSE2 on bootstrap since Factor now requires it
Joe Groff [Thu, 13 May 2010 21:41:40 +0000 (14:41 -0700)]
Merge branch 'master' of factorcode.org:/git/factor
Joe Groff [Thu, 13 May 2010 21:30:19 +0000 (14:30 -0700)]
math: declare quotation stack effects on each-integer, times, find-integer, all-integers?
Slava Pestov [Thu, 13 May 2010 05:46:58 +0000 (01:46 -0400)]
compiler: small fixes and cleanups
Joe Groff [Thu, 13 May 2010 01:07:33 +0000 (18:07 -0700)]
x11: convert XSupportsLocale return value to factor bool before testing (bug reported by ceninan)
Slava Pestov [Wed, 12 May 2010 22:07:11 +0000 (18:07 -0400)]
compiler.cfg.builder.alien: fix for x86-32
Joe Groff [Wed, 12 May 2010 20:49:02 +0000 (13:49 -0700)]
Merge branch 'master' of factorcode.org:/git/factor
Joe Groff [Wed, 12 May 2010 20:48:44 +0000 (13:48 -0700)]
new cuda.types vocab containing CUDA vector types (int2, float4, etc.) with CUDA alignment
Slava Pestov [Wed, 12 May 2010 06:09:11 +0000 (02:09 -0400)]
Refactor x86-32 stack cleanup logic
Slava Pestov [Wed, 12 May 2010 05:40:41 +0000 (01:40 -0400)]
Temporary fixes for x86-32 until FFI boxing is rewritten
Slava Pestov [Wed, 12 May 2010 03:24:25 +0000 (23:24 -0400)]
vm: fix longlong accessors
Slava Pestov [Wed, 12 May 2010 03:23:41 +0000 (23:23 -0400)]
Eliminate compiler.alien
Slava Pestov [Wed, 12 May 2010 02:29:46 +0000 (22:29 -0400)]
cpu.x86.64: fix typo
Slava Pestov [Wed, 12 May 2010 02:26:18 +0000 (22:26 -0400)]
Move flatten-c-type to death row so that it can be executed in part 5
Slava Pestov [Tue, 11 May 2010 23:29:50 +0000 (19:29 -0400)]
vm: fix compile error
Slava Pestov [Tue, 11 May 2010 23:11:31 +0000 (19:11 -0400)]
FFI rewrite part 4: parameter and return value unboxing redesign
Slava Pestov [Tue, 11 May 2010 00:50:28 +0000 (20:50 -0400)]
cuda.utils: fix load error
Slava Pestov [Mon, 10 May 2010 22:50:10 +0000 (17:50 -0500)]
cpu.ppc: fixes
Joe Groff [Mon, 10 May 2010 22:07:00 +0000 (15:07 -0700)]
Merge branch 'master' of factorcode.org:/git/factor
Joe Groff [Mon, 10 May 2010 22:06:15 +0000 (15:06 -0700)]
cuda.utils: add "sync-context" word
Slava Pestov [Mon, 10 May 2010 06:21:23 +0000 (02:21 -0400)]
cpu.ppc: update for recent changes -- untested
Slava Pestov [Mon, 10 May 2010 05:38:34 +0000 (01:38 -0400)]
compiler: simplify ##unbox-small/large-struct by emitting an ##unbox-any-c-ptr first
Slava Pestov [Mon, 10 May 2010 04:45:31 +0000 (00:45 -0400)]
sequences: remove unused 2reverse-each word
Slava Pestov [Mon, 10 May 2010 04:45:07 +0000 (00:45 -0400)]
FFI rewrite part 3: eliminate ##push-context-stack and ##pop-context-stack
Slava Pestov [Mon, 10 May 2010 04:42:03 +0000 (00:42 -0400)]
compiler.codegen.fixup: remove unused variable
Slava Pestov [Mon, 10 May 2010 04:24:13 +0000 (00:24 -0400)]
compiler.graphviz: fix load error
Slava Pestov [Mon, 10 May 2010 03:36:57 +0000 (23:36 -0400)]
cpu.x86.32: fix load error
Slava Pestov [Mon, 10 May 2010 03:25:46 +0000 (23:25 -0400)]
FFI rewrite part 2: use ##peek and ##replace instructions to access stack
Slava Pestov [Mon, 10 May 2010 02:15:14 +0000 (22:15 -0400)]
cpu.x86.64.unix: fix load error
Slava Pestov [Mon, 10 May 2010 01:55:21 +0000 (21:55 -0400)]
cpu.x86.32: fix load error
Slava Pestov [Mon, 10 May 2010 01:36:52 +0000 (21:36 -0400)]
FFI rewrite part 1: split up ##alien-invoke and friends into smaller instructions
Joe Groff [Sat, 8 May 2010 01:15:37 +0000 (18:15 -0700)]
cuda.devices: fix distribute-jobs to look only at width of maxThreadsDim
Joe Groff [Sat, 8 May 2010 01:09:21 +0000 (18:09 -0700)]
cuda.devices: have distribute-jobs construct a launcher with the calculated grid-dim, block-dim, and shared-size
Joe Groff [Sat, 8 May 2010 01:03:30 +0000 (18:03 -0700)]
remove cuda.constants; it's better to get those values from device-properties
Joe Groff [Sat, 8 May 2010 01:02:42 +0000 (18:02 -0700)]
cuda.devices: "distribute-jobs" utility word that determines grid and block size for a job based on per-thread shared memory requirements and max block size
Joe Groff [Sat, 8 May 2010 00:18:08 +0000 (17:18 -0700)]
cuda.utils: foo <c-object> ... foo memory>struct == foo <struct>
Joe Groff [Sat, 8 May 2010 00:16:28 +0000 (17:16 -0700)]
cdua.devices: remove redundant init-cuda calls
Slava Pestov [Fri, 7 May 2010 22:26:16 +0000 (18:26 -0400)]
ui.backend.windows: faster wheel mouse scroll rate
Slava Pestov [Fri, 7 May 2010 22:26:00 +0000 (18:26 -0400)]
compiler.cfg: add ##load-float instruction for single precision floating point constants
Slava Pestov [Fri, 7 May 2010 22:22:35 +0000 (18:22 -0400)]
compiler.cfg.linear-scan: clean up clobber-insn handling
Slava Pestov [Thu, 6 May 2010 21:41:00 +0000 (17:41 -0400)]
Load debugger.{unix,windows} later on in bootstrap, because lopading it before the optimizing compiler incurs a long delay on Windows as all the WinAPI bindings get loaded and parsed
Slava Pestov [Wed, 5 May 2010 20:52:54 +0000 (16:52 -0400)]
Language change: tuple slot setter words with stack effect ( value object -- ) are now named FOO<< instead of (>>FOO)
Joe Groff [Thu, 6 May 2010 06:26:54 +0000 (23:26 -0700)]
windows.directx.dinput: redo constant generation yet again to get compile time under control
Joe Groff [Thu, 6 May 2010 05:32:45 +0000 (22:32 -0700)]
silence more msvc warnings
Joe Groff [Thu, 6 May 2010 05:26:09 +0000 (22:26 -0700)]
define _CRT_SECURE_NO_WARNINGS to prevent MSVC from pushing its stupid nonstandard forms of standard C functions
Joe Groff [Thu, 6 May 2010 05:22:13 +0000 (22:22 -0700)]
quiet some data conversions warnings raised by msvc
Joe Groff [Thu, 6 May 2010 02:07:39 +0000 (19:07 -0700)]
cuda: fix using single integer as grid size
Joe Groff [Thu, 6 May 2010 00:38:59 +0000 (17:38 -0700)]
cuda.memory: dependencies
Joe Groff [Thu, 6 May 2010 00:38:37 +0000 (17:38 -0700)]
cuda: rearrange function-launcher grid and block dim to be in the right order
Joe Groff [Wed, 5 May 2010 23:36:12 +0000 (16:36 -0700)]
Merge branch 'master' of git://factorcode.org/git/factor into cuda-changes
Doug Coleman [Wed, 5 May 2010 07:06:03 +0000 (02:06 -0500)]
Fix typo in between? stack effect docs
Slava Pestov [Wed, 5 May 2010 17:37:25 +0000 (13:37 -0400)]
cpu.ppc: add missing cases to ##load/store-memory instructions
Slava Pestov [Wed, 5 May 2010 17:23:00 +0000 (13:23 -0400)]
compiler.cfg.intrinsics: may as well use cell-size stores instead of 32-bit stores when initializing byte arrays
Slava Pestov [Wed, 5 May 2010 17:17:20 +0000 (13:17 -0400)]
compiler.cfg.intrinsics.allot: fix <byte-array> intrinsic for real. Don't ever check in code without testing it
Daniel Ehrenberg [Wed, 5 May 2010 13:54:47 +0000 (08:54 -0500)]
Merge branch 'master' of git://factorcode.org/git/factor into propagation
Slava Pestov [Wed, 5 May 2010 06:54:48 +0000 (02:54 -0400)]
cpu.x86.32: fix load error
Slava Pestov [Wed, 5 May 2010 06:03:53 +0000 (02:03 -0400)]
compiler.cfg.intrinsics.allot: <byte-array> intrinsic was writing past the end of the array and this was causing problems for scheduling
Joe Groff [Wed, 5 May 2010 05:21:21 +0000 (22:21 -0700)]
Merge branch 'master' of factorcode.org:/git/factor into cuda-changes
Slava Pestov [Wed, 5 May 2010 05:13:45 +0000 (01:13 -0400)]
More FFI cleanups
Slava Pestov [Wed, 5 May 2010 04:44:21 +0000 (00:44 -0400)]
functors: fix unit test
Joe Groff [Wed, 5 May 2010 02:49:44 +0000 (19:49 -0700)]
add cuda.constants vocab
Joe Groff [Wed, 5 May 2010 02:46:53 +0000 (19:46 -0700)]
cuda.memory, cuda.syntax: inline some more words, add cuda-malloc-type word that allocates n*heap-size bytes of memory
Joe Groff [Wed, 5 May 2010 00:34:31 +0000 (17:34 -0700)]
Merge branch 'master' of factorcode.org:/git/factor into cuda-changes
Slava Pestov [Tue, 4 May 2010 23:33:46 +0000 (19:33 -0400)]
alien.c-types: cleanup
Slava Pestov [Tue, 4 May 2010 23:08:01 +0000 (19:08 -0400)]
bootstrap.compiler.timing: update
Daniel Ehrenberg [Tue, 4 May 2010 23:11:02 +0000 (18:11 -0500)]
Merge branch 'master' of git://factorcode.org/git/factor
Daniel Ehrenberg [Tue, 4 May 2010 23:10:34 +0000 (18:10 -0500)]
Cleaning up trees code a little bit
Slava Pestov [Tue, 4 May 2010 22:21:12 +0000 (18:21 -0400)]
mason.release.tidy: don't die if file doesn't exist
Slava Pestov [Tue, 4 May 2010 21:57:53 +0000 (17:57 -0400)]
vm: fix field order in zone struct
Daniel Ehrenberg [Sun, 25 Apr 2010 18:05:02 +0000 (13:05 -0500)]
Fixing bug in dependency tracking for guarded method inlining
Daniel Ehrenberg [Sun, 25 Apr 2010 07:56:23 +0000 (02:56 -0500)]
Bug fix in method inlining
Daniel Ehrenberg [Sun, 25 Apr 2010 07:18:45 +0000 (02:18 -0500)]
Removing hack with >fixnum, now subsumed by a more general optimization
Daniel Ehrenberg [Sun, 25 Apr 2010 07:12:23 +0000 (02:12 -0500)]
Adding dependency checking and proper handing of dispatch# to the new method inlining
Daniel Ehrenberg [Sun, 25 Apr 2010 00:46:12 +0000 (19:46 -0500)]
Improving method inlining change, but there's still a bug
Daniel Ehrenberg [Tue, 20 Apr 2010 05:39:08 +0000 (00:39 -0500)]
Changes to compiler.tree.propagation.inlining
Daniel Ehrenberg [Mon, 19 Apr 2010 21:57:17 +0000 (16:57 -0500)]
Fixing another bug in the method inlining improvement
Daniel Ehrenberg [Mon, 19 Apr 2010 21:37:35 +0000 (16:37 -0500)]
Fixing optimization for improved method inlining
Daniel Ehrenberg [Mon, 19 Apr 2010 21:01:09 +0000 (16:01 -0500)]
splitting generic word callsites when only one method is applicable