Joe Groff
|
cee5862b69
|
update a bunch of alien-callbacks and alien-indirects to use c-type words
|
2009-10-21 21:10:11 -05:00 |
|
Joe Groff
|
4dd44a007a
|
update some naked alien-invokes to use c-type words
|
2009-10-21 18:44:00 -05:00 |
|
Joe Groff
|
a2976083be
|
generate unsigned vector comparison fallbacks using min/max or xor/signed compare
|
2009-10-20 22:30:57 -05:00 |
|
Joe Groff
|
a0a816e88a
|
generate better fallback code for vmin/vmax intrinsics
|
2009-10-20 19:22:38 -05:00 |
|
Joe Groff
|
d570e58374
|
Merge branch 'master' of git://factorcode.org/git/factor
|
2009-10-20 17:42:29 -05:00 |
|
Joe Groff
|
218d5c7c07
|
value numbering rewrite rules for vector ops to convert "not and" to "andn" and "not andn" to "and"
|
2009-10-20 17:29:01 -05:00 |
Slava Pestov
|
a71212f9c4
|
cpu.x86.32: only create 16-byte parameter area if the word calls into the VM
|
2009-10-20 05:02:42 -05:00 |
|
Slava Pestov
|
9599d86ec5
|
compiler: FFI is now slightly more efficient when unboxing parameters, only changes data stack height once
|
2009-10-20 04:31:48 -05:00 |
Slava Pestov
|
0a3029d9f2
|
compiler: on PPC, ANDI, ORI and XORI instructions take an unsigned 16-bit immediate, unlike ADDI, SUBI and MULLI which take a signed 16-bit immediate. The code generator was not aware of this, and so for example '[ >fixnum -16 bitand ]' would generate incorrect code. Split up small-enough? hook into immediate-arithmetic? and immediate-bitwise? and update value numbering to be aware of this. Fixes classes.struct bitfields test failure
|
2009-10-19 04:58:29 -05:00 |
|
Joe Groff
|
bff46b80a8
|
fix unsigned vector unpack
|
2009-10-16 14:25:33 -05:00 |
|
Slava Pestov
|
ee8a112c44
|
compiler.cfg: fix unit tests
|
2009-10-15 06:01:46 -05:00 |
|
Slava Pestov
|
22e79e8495
|
compiler: tweak ##write-barrier-imm
|
2009-10-15 02:40:23 -05:00 |
|
Slava Pestov
|
10ad5cad53
|
Working on adding support for the new write barrier to optimized code
|
2009-10-14 02:06:01 -05:00 |
|
Joe Groff
|
2577ab83a6
|
only emit ##alien-vector/##set-alien-vector insns if the rep is available
|
2009-10-10 12:53:10 -05:00 |
|
Joe Groff
|
61befc8bb1
|
have vshuffle accept simd-128 variable byte shuffles
|
2009-10-10 11:30:11 -05:00 |
|
Joe Groff
|
5158a12d32
|
rename ##shuffle-vector to ##shuffle-vector-imm, and add a new ##shuffle-vector for dynamic shuffles. have vshuffle use ##shuffle-vector to do word and byte shuffles on x86
|
2009-10-09 21:26:27 -05:00 |
|
Joe Groff
|
2bfcd7ed81
|
generate better code for vabs when instruction isn't available instead of using software fallback (-0.0 andn for floats, x > 0 ? x : -x for signed ints, nop for unsigned ints)
|
2009-10-09 14:24:55 -05:00 |
|
Joe Groff
|
4972fbcbc9
|
implement vneg as an intrinsic in terms of load -0, subtract
|
2009-10-09 13:16:39 -05:00 |
Daniel Ehrenberg
|
ab919bab96
|
Doing constant folding on ##neg and ##not in value numbering
|
2009-10-08 01:57:54 -05:00 |
|
Joe Groff
|
98836a9e2e
|
break vector compare intrinsics into %compare, %or, and %not instructions that map directly to cpu instructions
|
2009-10-07 15:27:03 -05:00 |
|
Joe Groff
|
43b51ef2eb
|
decompose %unpack-vector-head/tail into %compare-vector/%merge-vector-head/tail or %tail>head-vector/%unpack-vector-head insns when there isn't an actual unpack insn; get rid of fake x86 implementations
|
2009-10-07 14:09:46 -05:00 |
|
Joe Groff
|
0dfeb74176
|
typo in convert-to-fill-vector?
|
2009-10-07 12:53:10 -05:00 |
|
Joe Groff
|
2d79729193
|
fix ##load-constant/##scalar>vector folding when constant is a fixnum
|
2009-10-07 12:46:08 -05:00 |
|
Joe Groff
|
aee85401af
|
convert all-ones vector ##load-reference/##load-constant to a ##fill-vector insn
|
2009-10-07 12:35:21 -05:00 |
|
Joe Groff
|
a13e75f4f4
|
don't generate a ##not-vector instruction if the cpu doesn't have one; instead, fall back to a ##fill-vector/##xor-vector combo. get rid of pretend %not-vector in cpu.x86
|
2009-10-07 11:59:36 -05:00 |
|
Joe Groff
|
444624e79f
|
fix x86 %unpack-vector insns
|
2009-10-06 20:38:51 -05:00 |
|
Joe Groff
|
96cb1d5038
|
glue conversion intrinsics to instructions
|
2009-10-06 20:13:38 -05:00 |
|
Joe Groff
|
c0423d2f3f
|
Merge branch 'master' of git://factorcode.org/git/factor
|
2009-10-05 22:34:41 -05:00 |
|
Joe Groff
|
84ecb1266d
|
add insns for vector pack, unpack, integer>float, and float>integer
|
2009-10-05 22:34:14 -05:00 |
|
Slava Pestov
|
a6953ec9ee
|
Merge branch 'master' of git://factorcode.org/git/factor
|
2009-10-05 18:22:09 -05:00 |
|
Slava Pestov
|
ee7cfeba7b
|
compiler.cfg.linear-scan: fix unit test
|
2009-10-05 18:21:58 -05:00 |
|
Joe Groff
|
3e44e0bb02
|
rename SIMD vmerge and kids to (vmerge), make new vmerge more generally useful
|
2009-10-05 17:55:39 -05:00 |
|
Slava Pestov
|
931107397c
|
compiler.cfg: remove _gc instruction, it doesn't need to exist, and change GC checks to ensure that the right amount of space is available instead of blindly checking for 1Kb
|
2009-10-05 05:27:49 -05:00 |
|
Joe Groff
|
dca9d3e535
|
add %merge-vector-head and %merge-vector-tail instructions to back vmerge
|
2009-10-03 21:48:53 -05:00 |
|
Joe Groff
|
335df20713
|
add intrinsics for v<=, v<, v=, v>, v>=, vunordered?
|
2009-10-03 11:29:34 -05:00 |
|
Joe Groff
|
b1ec36a324
|
extend x86 %compare-vector to cover all comparison codes, sometimes stupidly for now
|
2009-10-02 23:19:56 -05:00 |
|
Joe Groff
|
e2e75c6b3a
|
add intrinsic for vnot/vbitnot
|
2009-10-02 20:04:28 -05:00 |
|
Joe Groff
|
e153d544eb
|
implement vand, vor, vandn, and vxor as bitwise intrinsics for simd types
|
2009-10-02 14:17:01 -05:00 |
|
Slava Pestov
|
d47e99dbde
|
Fix compiler.cfg.value-numbering and math.matrices.simd unit tests
|
2009-10-02 02:32:35 -05:00 |
|
Slava Pestov
|
901e0709e6
|
Merge branch 'master' of git://factorcode.org/git/factor into improved-aa
|
2009-10-02 00:21:44 -05:00 |
|
Slava Pestov
|
10a49a4613
|
compiler.cfg.alias-analysis: fix ##compare where operands are copies
|
2009-10-01 23:54:19 -05:00 |
|
Slava Pestov
|
231a2d1b71
|
Fix alias analysis of a ##compare whose input is another ##compare
|
2009-10-01 23:20:05 -05:00 |
|
Joe Groff
|
9d424a1092
|
Merge branch 'master' of git://factorcode.org/git/factor
Conflicts:
basis/compiler/codegen/codegen.factor
|
2009-10-01 23:14:16 -05:00 |
|
Slava Pestov
|
d3acd0a530
|
compiler.cfg.alias-analysis: optimize ##vm-field-ptr and ##alien-global instructions, and optimize out ##compare between values of different alias classes; this optimizes '[ [ >float ] bi@ [ + ] [ - ] 2bi eq? ]' down to an o-op and removes boxing from '[ [ >float ] bi@ [ + ] [ - ] 2bi = ]'
|
2009-10-01 23:03:17 -05:00 |
|
Joe Groff
|
7b13fa4283
|
fold test-vector/branch sequences into a test-vector-branch instruction
|
2009-10-01 19:53:30 -05:00 |
|
Slava Pestov
|
6740956162
|
compiler.cfg: don't unbox the same value more than once per basic block
|
2009-10-01 19:41:23 -05:00 |
|
Slava Pestov
|
e6ff877362
|
compiler.cfg.instructions: remove ##box-float, ##unbox-float, ##box-vector, ##unbox-vector since they can be expressed in terms of ##alien-double, ##set-alien-double, ##alien-vector, ##set-alien-vector, and ##allot
|
2009-10-01 18:07:50 -05:00 |
|
Joe Groff
|
228ad950bb
|
%test-vector instruction for vany?, vall?, vnone?
|
2009-10-01 15:35:38 -05:00 |
|
Joe Groff
|
94070c11aa
|
%compare-vector instruction (only does v= for now)
|
2009-10-01 14:31:37 -05:00 |
|
Joe Groff
|
3ba79be651
|
Revert "add a %blend-vector intrinsic for v?"
This reverts commit 21e4b28b67.
|
2009-09-30 23:40:37 -05:00 |