Commit Graph

722 Commits

Author SHA1 Message Date
dzaima
70705f9bfe move bmi.singeli into base x86 include 2025-04-27 19:07:39 +03:00
dzaima
98f1c97653 is_pow2 helper 2025-04-24 18:31:25 +03:00
dzaima
0ec37303ea use specified-output-type fold_addw 2025-04-24 00:26:49 +03:00
dzaima
3d52d14101 use vecfold.singeli's x86 fold_addw in count.singeli 2025-04-24 00:26:49 +03:00
dzaima
5df3712748 allow specifying result type of fold_addw, x86 impls for it 2025-04-24 00:26:49 +03:00
dzaima
7237ad1abb more Singeli renames: no more floor,ceil,abs,adds,subs,vshl 2025-04-24 00:23:50 +03:00
dzaima
2202726005 rename isvec→vect and isprim→ptrt 2025-04-24 00:23:50 +03:00
dzaima
6eb454f859 get rid of the single isptr usage 2025-04-24 00:23:50 +03:00
dzaima
c1662c83de move x86 mul32 and mulh defs to x86.singeli 2025-04-24 00:23:50 +03:00
dzaima
e46bbf398e 512-bit & arbitrary-width vector check helpers 2025-04-24 00:23:50 +03:00
dzaima
64a71688fe use util/kind checks instead of custom isreg/istype/istup 2025-04-24 00:23:50 +03:00
dzaima
ee1668c7ea use __min & __max directly instead of custom min/max
might make sense to bring min/max back later as aliases, but the current situation of min!=__min was stupid
2025-04-24 00:21:48 +03:00
dzaima
8beff254e2 x86 saturating multiply sum defs 2025-04-24 00:00:32 +03:00
dzaima
44ee417d40 use pattern matching for NEON vfold cases 2025-04-24 00:00:32 +03:00
dzaima
732f52630a use AVX-512 integer narrow if available 2025-04-24 00:00:32 +03:00
dzaima
528fea645a nicer store-low in bins.singeli
also means it doesn't break on width>256
2025-04-23 22:58:58 +03:00
dzaima
78359afc71 move x86 int widen into x86.singeli
+ AVX-512 defs for it & (undef|zero)_promote; + add extract{V, x:V0, k}
2025-04-03 03:53:31 +03:00
dzaima
dbe97cf0e0 move some basic things into x86.singeli 2025-04-03 03:53:25 +03:00
dzaima
5928bd91bc note problematic latency of dpbusd/dpwssd 2025-04-03 00:37:11 +03:00
dzaima
63b7fab72c better w+x & w-x overflow checks 2025-03-26 04:16:38 +02:00
dzaima
68a969b417 get clang to do less horrible codegen for index calculations
the do-while loop isn't actually necessary, just moving the unr offset is enough, but I'm doing it anyway
2025-03-24 18:12:55 +02:00
dzaima
ad73e075e0 unroll bitwise arith 2025-03-24 18:12:55 +02:00
dzaima
60b848f1b7 don't require strict subset type for trunc 2025-03-22 19:09:51 +02:00
dzaima
c7632c23a2 actually, can define the x86 sadbw-using vfold for i8
it's unsigned, but that doesn't matter when just taking the low 8 bits
2025-03-22 18:21:09 +02:00
dzaima
ae29c4ad45 fix usz=32 build 2025-03-21 18:08:12 +02:00
dzaima
609f3418d8 use offset loads instead of vec_merge_shift_right for groupstat 2025-03-21 06:01:27 +02:00
dzaima
e2b4dbcf20 use any_hom / all_hom for (any|all)_accumulator 2025-03-21 06:01:27 +02:00
dzaima
05d87dd7df move groupstat to fancy accumulator 2025-03-21 06:01:26 +02:00
dzaima
97260e0051 special-cased x86 SSE2 i8 min/max accumulator 2025-03-21 04:32:26 +02:00
dzaima
f503880cc3 use fancy accumulator for getRange 2025-03-21 04:32:26 +02:00
dzaima
e1b62b95b6 fancy Singeli accumulator; use for integer sum 2025-03-21 04:32:26 +02:00
dzaima
3b1239d499 support requesting a flush every k iterations in for_mu 2025-03-21 04:32:26 +02:00
dzaima
9c2ea18e22 use x86 sad for vfold
not used by anything, but it's here
2025-03-21 04:32:26 +02:00
dzaima
6c6f6e4013 add x86 sad & madd intrinsics 2025-03-21 04:32:26 +02:00
dzaima
9876d1d622 reduce Singeli-generated C noise on constant broadcasts 2025-03-21 04:32:26 +02:00
dzaima
4a5a1fa35b a couple more baseline utilities
assert_hom to be expanded upon later
2025-03-21 01:05:53 +02:00
dzaima
87be577ff7 actually use defined is in for_mu 2025-03-20 20:27:26 +02:00
dzaima
0c2fb1a267 a bit of cleanup
still from the long-ago days when symbols couldn't be assignment targets directly
2025-03-19 01:05:22 +02:00
dzaima
577b09a062 introduce x86.singeli 2025-03-19 01:05:15 +02:00
Marshall Lochbaum
233c88120d Singeli function for initial group statistics pass since auto-vectorization isn't great 2025-03-13 14:21:38 -04:00
dzaima
d48194c3c8 fix comment 2025-03-12 17:20:18 +02:00
dzaima
7a412523dd better SSE2 float getRange 2025-03-12 16:38:20 +02:00
dzaima
a013eb094e select_fn doesn't require SIMD anymore
it's not used too much without SIMD extensions available, but it is used a bit
2025-03-12 04:20:53 +02:00
dzaima
d2ca600e02 actually make try_first_hom return things 2025-03-12 00:20:49 +02:00
dzaima
85ef1038b8 fix usz=32 build 2025-03-11 23:41:11 +02:00
Marshall Lochbaum
b69a2c2469 Move swtab outside of its function; it was being compiled as a dynamic array 2025-03-09 08:44:42 -04:00
Marshall Lochbaum
e898a81d68 Use a possibly-overlapping word and constant-length loop in k/bool, like ⊣` 2025-03-09 08:44:42 -04:00
Marshall Lochbaum
302c2f926f AVX2 k/bool for odd 64<k<256 using shift-based masks 2025-03-09 08:44:42 -04:00
dzaima
bd942894bf fix missing usage of unaligned store 2025-03-06 06:01:44 +02:00
dzaima
d5facc33ad some Singeli cleanup
shuf_ind's staying as such I guess; make it more clear when debug.singeli is left included
2025-03-06 05:32:22 +02:00