4d5a4222b6
More SIMD work
...
- Rename SIMD types and register representations: <type>-<count> rather than <count><type>-array
- Make a functor to define 256-bit vector types, use it to define float-8 type
- Make SIMD instructions pure-insns so that they participate in value numbering
2009-09-03 20:58:56 -05:00
ff8c70dbe0
Initial implementation of SSE vector intrinsics:
...
- cpu.architecture: add SSE vector representations
- compiler.cfg.intrinsics.alien: remove an attempt at optimization that value numbering handles now
- compiler.cfg.representations: support instructions where the representation is set in the 'rep' slot, and support conversions between single and double floats
- alien-float, set-alien-float now use the single float representation, and the conversion is implicit; this fixes a long-standing bug where a register could get clobbered because of how %set-alien-float was defined on x86
- math.vectors.specialization: add support for SIMD specialization (where the vector word's body is replaced by another quotation), also specialize the 'sum' word
- math.vectors.simd: 4float-array, 2double-array, 4double-array types, and specializers for the math.vectors words
2009-09-03 02:33:07 -05:00
85a2bfab6c
compiler: eliminate boilerplate by centralizing info in declarative INSN: syntax
2009-09-02 06:22:37 -05:00
9595be4bf9
%box-displaced-alien: fix clobberage found by Doug
2009-08-30 05:11:08 -05:00
0db01f6d5f
compiler.cfg.linear-scan now supports partial sync-points where all registers are spilled; taking advantage of this, there are new trigonometric intrinsics which yield a 2x performance boost on benchmark.struct-arrays and a 25% boost on benchmark.partial-sums
2009-08-30 04:52:01 -05:00
2bb6293217
compiler: add fixnum-min/max intrinsics; ~10% speedup on benchmark.yuv-to-rgb
2009-08-28 19:02:59 -05:00
d957ae4e44
Performance improvements to make struct-arrays benchmark faster
...
- improved optimization of ##unbox-any-c-ptr on ##box-displaced-alien; convert it to ##unbox-c-ptr where possible using class info stored in the ##bda instruction
- make fcos, fsin, etc inline again; everything in math.libm inline again, except for fsqrt which is an intrinsic
- convert min and max on floats to float-min and float-max
- make min and max not inline, so that the above can work
- struct-arrays: rice a bit so that more fixnums come up
2009-08-28 05:21:16 -05:00
f662e6403a
compiler: new inline intrinsic for <displaced-alien> where the inputs have known types; value numbering now eliminates unnecessary allocation of displaced aliens if the result is immediately unboxed again
2009-08-27 00:06:19 -05:00
0df8aadce2
cpu.x86: use SQRTSD instruction for math.libm:fsqrt word
2009-08-25 23:22:15 -05:00
1961b4da16
next-fastcall-param word was not being called; on x86 its equivalent to inc but on ppc there is more logic, this fixes FFI on PowerPC
2009-08-21 20:15:19 -05:00
cc5476c823
_gc instruction doesn't need slot to hold GC root area size, since that's just tagged-values>> length
2009-08-09 03:08:13 -05:00
725280d424
Split off the notion of a register representation from a register class
2009-08-07 17:44:50 -05:00
45770c6250
Move a bunch of GC check generation logic to platform-independent side
2009-07-30 21:28:27 -05:00
be363d1a5b
compiler.cfg: Get inline GC checks working again, using a dataflow analysis to compute uninitialized stack locations in compiler.cfg.stacks.uninitialized. Re-enable intrinsics which use inline allocation
2009-07-30 09:19:44 -05:00
74766d1ccd
compiler.cfg.linear-scan.assignment: modifies instructions in place instead of storing a registers assoc for further compile-time performance improvement
2009-07-29 06:36:14 -05:00
31555b05cf
compiler.codegen: collect instruction statistics
2009-07-27 03:58:15 -05:00
44bcd258f6
Insert _loop-entry in linearization pass instead of in CFG builder, so that optimizations don't have to worry about it
2009-07-22 06:05:17 -05:00
3fb4fc1bde
Improve code generation for shift word: add intrinsics for fixnum-shift-fast in the case where the shift count is not constant, transform 1 swap shift into a more overflow check with open-coded fast case, transform bitand into fixnum-bitand in more cases
2009-07-16 23:50:48 -05:00
e76dce8aff
Overflowing fixnum intrinsics now expand into several CFG nodes. This speeds up the common case since only the uncommon case is now a stack syncpoint
2009-07-16 18:29:40 -05:00
0402790001
compiler.cfg Remove height tracking for ##call instructions, wire in ##no-tco instruction
2009-06-30 21:21:46 -05:00
d0f6a7d048
Split up compiler.cfg.linear-scan.allocation into a number of sub-vocabularies; start work on compiler.cfg.linear-scan.resolve; start work on inactive interval splitting
2009-06-11 17:55:14 -05:00
9ad9600764
Remove %dispatch-label since its tehe same on all platforms; fix %gc on PowerPC
2009-06-07 21:46:28 -05:00
2d231f066a
GC checks now save and restore registers
2009-06-02 18:23:47 -05:00
b389dcf441
Redo compiler.codegen.fixup and get %dispatch to work
2009-06-01 02:32:36 -05:00
fc152ef210
Various improvements aimed at getting local optimization regressions fixed:
...
- Rename _gc to ##gc
- Absolute labels are now supported
- Generate _dispatch-label
2009-05-31 23:28:08 -05:00
76d74c16af
Fixing various bugs; alias analysis wasn't handling ##phi nodes, stack analysis incorrectly handled height-changing back edges and ##fixnum-*, clean up ##dispatch generation
2009-05-29 01:39:14 -05:00
a366909c25
Removing slip usage from basis
2009-05-10 15:28:22 -05:00
c93d876075
Better separation of concerns: cpu.{x86,ppc}.assembler no longer depends on compiler.codegen.fixup and cpu.architecture. Rename rt-xt-direct to rt-xt-pic to better explain its purpose
2009-05-06 16:14:53 -05:00
fb5b48c71a
Fix some test failures
2009-05-04 09:44:26 -05:00
515c619202
Non-optimizing compiler now open-codes megamorphic dispatch fast path
2009-04-30 03:37:07 -05:00
5649cc7a0a
Compiler warnings are no more
2009-04-23 22:17:25 -05:00
c4be8bd65a
Refactor error code so that error types are registered; listener prints a summary of errors
2009-04-12 16:08:46 -05:00
0a7485190b
compile-error-type => source-error-type; make test failures global
2009-04-10 03:52:12 -05:00
39b53817b9
Small size reduction for deployed images
2009-03-31 08:15:27 -05:00
8b7fcce399
update code for usages of add-library
2009-03-25 23:00:19 -05:00
7f4c967ace
More compact relocation layout
2009-03-19 20:02:43 -05:00
be4fb1e7d9
Move call( and execute( to core
2009-03-16 20:11:36 -05:00
38c6be94d4
Fixing some stack comments
2009-03-16 06:17:18 -05:00
7bb0e78314
Add support for C99 complex float and complex double types to FFI
...
They are named complex-float and complex-double in the Factor world
2009-02-06 04:02:00 -06:00
3c408342ef
renaming: contain? -> any?, deep-contains? -> deep-any?, pad-left -> pad-head, pad-right -> pad-tail, trim-left -> trim-head, trim-right -> trim-tail
2009-01-29 22:19:07 -06:00
44bfff7c7b
Rename ##load-indirect to ##load-reference since this is more descriptive; value numbering doesn't assign expressions to ##load-reference nodes since this would end up folding literals which were eq? but not =
2009-01-29 01:44:58 -06:00
4204fd495f
Move formatted stream output words to io.styles; they didn't belong in core, and this reduces deployed image size
2009-01-13 17:12:43 -06:00
0021cde743
Fix more compile errors
2008-12-15 22:21:56 -06:00
daf2857d7b
define-inline now takes an effect parameter
...
C: now sets the stack effect of the new word to slot names
add new basis/constants vocab; CONSTANT: foo bar is equivalent to : foo bar ; inline
(going to fold CONSTANT:, ALIAS: and qualified into core soon perhaps)
tighten the screws: inline words now need a stack declaration too
update some usages of inline words
2008-12-15 19:44:56 -06:00
8a8f0c925c
Use BSR instruction to implement fixnum-log2 intrinsic
2008-12-06 15:31:17 -06:00
a56d480aa6
Various optimizations leading to a 10% speedup on compiling empty EBNF parser:
...
- open-code getenv primitive
- inline tuple predicates in finalization
- faster partial dispatch
- faster built-in type predicates
- faster tuple predicates
- faster lo-tag dispatch
- compile V{ } clone and H{ } clone more efficiently
- add fixnum fast-path to =; avoid indirect branch if two fixnums not eq
- faster >alist on hashtables
2008-12-06 09:16:29 -06:00
f48653c47a
Fix compile error
2008-12-05 23:51:34 -06:00
e256846acd
Tweak string representation; high bit indicates if character has high bits in aux vector. Avoids memory access in common case. Split set-string-nth into two primitives; set-string-nth-fast is open-coded by optimizing compiler. 13% improvement on reverse-complement
2008-12-05 06:38:51 -06:00
6dce834d91
Get rid of some more >r/r> usages
2008-11-30 18:28:15 -06:00
e7f4563374
fixnum* intrinsic for x86
2008-11-30 07:26:49 -06:00
Slava Pestov
Slava Pestov
Slava Pestov
Slava Pestov
Doug Coleman
Doug Coleman