Commit Graph

  • 29b2b1c1f5 fix singeli-lsp -l path dzaima 2025-01-26 06:04:15 +0200
  • d506691690 fix 𝕨⍋𝕩 issues around NaNs dzaima 2025-02-01 20:55:04 +0200
  • 9be5bf45b4 fix signed offset having unsigned type dzaima 2025-01-26 00:06:54 +0200
  • d06200445b MAYBE_UNUSED dzaima 2025-01-25 00:18:25 +0200
  • dcbceccff8 fix integer wrapping in grade sum check dzaima 2025-01-24 03:36:01 +0200
  • 204ec21a7c more compact vector width check definitions dzaima 2025-01-23 03:21:55 +0200
  • 5f19f15556 NEON (hom|bit)(Any|All) improvements dzaima 2025-01-22 21:34:16 +0200
  • 196cc07454 better isCallable dzaima 2025-01-17 21:51:46 +0200
  • 8b0d0399e4 fix various F⌾(G⊸Specialized)x dzaima 2025-01-16 01:00:35 +0200
  • 64a3f44f3f
    Merge pull request #131 from vylsaz/utf16 dzaima 2025-01-16 22:08:19 +0200
  • 517300e567 Windows: static-method-only .c files to .h dzaima 2025-01-16 21:41:26 +0200
  • 7d928a3221 Use OsStr for path on both Windows and non-Windows systems vylsaz 2025-01-11 18:09:47 +0000
  • 24bd6735bc Windows: use _wfullpath for realpath vylsaz 2025-01-11 18:08:33 +0000
  • e12ca55ee0 Windows: use utf16 for file functions vylsaz 2025-01-11 05:46:39 +0000
  • 40609b2f5f Windows: sh: revert to using calloc() in reading threads vylsaz 2025-01-11 05:45:35 +0000
  • 67c2850e38 Windows: sh using utf16; simply argument quoting vylsaz 2025-01-07 03:02:16 +0000
  • 580f4a3a19 Windows: utf-16 util vylsaz 2025-01-07 03:00:59 +0000
  • 6bf8f4d3c7 Include packaging status in readme dzaima 2025-01-16 00:52:09 +0200
  • a02f195555 make native inverse bindings for ⋆⁼x & w⋆⁼x dzaima 2024-12-18 13:29:32 +0200
  • 1822602c34 cleanup from clang's --analyze dzaima 2024-12-18 13:03:48 +0200
  • d1a9669e76 fix •math.LCM on arrays dzaima 2024-12-13 07:37:08 +0200
  • 5ecd1858ef fix 1⌾(⟨0⟩⊸⊏)⟨0⟩ dzaima 2024-12-07 06:54:07 +0200
  • 86c7b47771 make •FFI properly error on non-list 𝕩 dzaima 2024-12-04 00:11:43 +0200
  • 162e6ff387 decG more dzaima 2024-12-04 00:10:13 +0200
  • 7a5f42a7f0 cleanup dzaima 2024-12-03 11:03:56 +0200
  • ed1fe6cc3e don't put needlessly absolute paths in Singeli compile commands dzaima 2024-12-03 11:03:34 +0200
  • 81abc80863 fast highrank⊸⊏˘ table dzaima 2024-12-02 21:53:45 +0200
  • 7509d0c238 don't shadow types dzaima 2024-11-30 20:22:58 +0200
  • d988a620fd native F⌾(a⊸⥊) dzaima 2024-11-26 18:08:15 +0200
  • 44e8c1e7fa expand feature list for Singeli compile commands dzaima 2024-11-23 16:04:52 +0200
  • 58f96abe8c fix error message on missing absolute base path dzaima 2024-11-23 06:19:39 +0200
  • 79cffc5011 don't use deprecated syntax dzaima 2024-11-22 12:48:45 +0200
  • 92641f26ba squeeze fns haven't been AVX2-specific for a while dzaima 2024-11-19 11:32:31 +0200
  • b143892f21
    Merge pull request #126 from mlochbaum/count dzaima 2024-11-22 00:55:09 +0200
  • 7df2f9ba4f a couple renames & simplifications dzaima 2024-11-21 13:20:36 +0200
  • f0223b8baf properly add 1<<31 dzaima 2024-11-21 23:02:51 +0200
  • 33ae8744f5 Fast skip ahead for blocks of equal values in i8 and i16 /⁼ Marshall Lochbaum 2024-11-17 10:16:48 -0500
  • 7baddcabc2 Better sum versus runs test for /⁼: account for per-element cost of run-based Marshall Lochbaum 2024-11-17 09:23:51 -0500
  • 342cadfc00 Check for or handle argument length >2⋆31 in /⁼ Marshall Lochbaum 2024-11-16 21:20:54 -0500
  • 194e8c3f22 Update /⁼ implementation comments Marshall Lochbaum 2024-11-16 20:57:47 -0500
  • 0bbb335893 Do general i8 and i16 /⁼ counts to i16 buffer, plus overflow list Marshall Lochbaum 2024-11-16 20:56:25 -0500
  • fb5ee179cb Squeeze for slash_im generic case instead of converting individual elements to numbers Marshall Lochbaum 2024-11-16 08:27:36 -0500
  • 96b198f643 Condition all the i8 to i32 /⁼ optimization on SINGELI_SIMD, for simplicity Marshall Lochbaum 2024-11-16 08:17:12 -0500
  • e6940e73d0 Fast /⁼ of sorted arguments using semi-sparse representation Marshall Lochbaum 2024-11-15 20:55:15 -0500
  • 11117fcc67 Version of count_i32_i32 without large blocks Marshall Lochbaum 2024-11-15 17:37:11 -0500
  • 3b103aadd0 Run-based i32 /⁼ (leaving out sum-based) Marshall Lochbaum 2024-11-14 22:18:59 -0500
  • 4aea668a6c Allow i16 /⁼ to jump to Singeli code after range checking Marshall Lochbaum 2024-11-14 15:07:20 -0500
  • 926e59483d More precise sparse check for i8 /⁼ result: previously stopped at 65 Marshall Lochbaum 2024-11-14 12:06:54 -0500
  • e24d0bac63 Extend SIMD counting to i16 in addition to i8 Marshall Lochbaum 2024-11-14 11:58:13 -0500
  • e681f3c09a Run-based 1-byte /⁼ implementation Marshall Lochbaum 2024-04-29 21:04:27 -0400
  • 092ba4167a fix talloc_realloc in VERIFY_TAIL dzaima 2024-11-17 22:18:33 +0200
  • f3094295dd Singeli compile commands dzaima 2024-11-12 10:14:22 +0200
  • 0a4925e84f move compile_commands.json to build/ dzaima 2024-11-12 03:11:50 +0200
  • b72fe465a2 fix /⁼1e4⥊1‿2 on NEON dzaima 2024-11-14 22:45:02 +0200
  • cbac419b4a fix ⍷2‿0⥊x dzaima 2024-11-06 22:29:32 +0200
  • d47cedb165 some cleanup dzaima 2024-11-06 22:24:25 +0200
  • 163853439e
    Merge pull request #123 from mlochbaum/modperm dzaima 2024-11-06 22:23:20 +0200
  • db16b8fa6e Fix dropped high bit in ⊣˝˘ due to signed rather than unsigned saturating pack Marshall Lochbaum 2024-11-06 14:52:18 -0500
  • 67d3fe3f81 Missing SSSE3 requirement for some uninterleave implementations Marshall Lochbaum 2024-11-06 14:51:16 -0500
  • 15df6afc7d Implementation comments for transpose with short height/width Marshall Lochbaum 2024-11-05 22:04:23 -0500
  • e75b63831f Clean up dispatching for the various transpose kernel methods Marshall Lochbaum 2024-11-05 21:39:15 -0500
  • 2d75c6c535 And manually vectorize uninterleave Marshall Lochbaum 2024-11-05 09:36:54 -0500
  • 32ad2e9953 Manually vectorized interleave function Marshall Lochbaum 2024-11-04 22:22:05 -0500
  • 980e4b2248 Short-height version of 16x16 i8 transpose kernel, using overlapped writes Marshall Lochbaum 2024-11-04 14:21:22 -0500
  • abb96cb18a Fixed-height transposes, more or less reverse of fixed-width Marshall Lochbaum 2024-11-03 20:20:31 -0500
  • 0ab2e485ca Width-4 transpose with shuffles Marshall Lochbaum 2024-11-02 11:36:10 -0400
  • 8bc37098dc Modular kernel transpose for odd*2 (so, just 6) Marshall Lochbaum 2024-11-01 21:31:50 -0400
  • 091f08c6cc Extend modular permutation transpose to any element size Marshall Lochbaum 2024-11-01 17:30:16 -0400
  • 780bfdfa0b Kernel transpose with overreading and skipped writes, for short rows Marshall Lochbaum 2024-10-31 19:16:50 -0400
  • b6e418ed5b Fix modular permutation example in comment Marshall Lochbaum 2024-10-30 15:09:56 -0400
  • d7b508ff3b Move select_cells_single (atom⊸⊏˘) to select.c and use for any singleton index Marshall Lochbaum 2024-10-29 22:38:06 -0400
  • 0bdc43cc0f Even number handling for modular permutation ⊣˝˘ Marshall Lochbaum 2024-10-29 16:17:15 -0400
  • 4c55eab740 Fast ⊣˝˘ for power-of-two widths, 4- and 8-byte elements Marshall Lochbaum 2024-10-28 20:39:02 -0400
  • 8d9d946ec5 Fast ⊣˝˘ with pack for power-of-two widths, 1- and 2-byte elements Marshall Lochbaum 2024-10-27 21:52:43 -0400
  • aefd2a4abd Fast 1-byte small odd width transpose with blends and modular permutation Marshall Lochbaum 2024-10-27 11:04:24 -0400
  • 5bdee886cc Extend kernel transpose from AVX2 only to SSE2, NEON Marshall Lochbaum 2024-10-26 20:08:56 -0400
  • 6e4859e17f Extend select_rows_byte to any architecture with (shuffle and) blend Marshall Lochbaum 2024-10-25 15:31:29 -0400
  • 4669418f1c Extend modular permutation ⊣˝˘ to multi-byte element types Marshall Lochbaum 2024-10-24 22:21:35 -0400
  • 9997c52e4c Initial 1-byte ⊣˝˘ with blends and modular permutation Marshall Lochbaum 2024-10-23 14:32:41 -0400
  • 18f594e829 Update Singeli submodule Marshall Lochbaum 2024-11-05 21:48:25 -0500
  • 0c6e32a24c fix si_select_cells_bit_lt64 signature dzaima 2024-10-23 21:53:48 +0300
  • 4c8b8519fa
    Merge pull request #122 from mlochbaum/vec-select dzaima 2024-10-23 21:51:22 +0300
  • bd6effe7ab restore dropped comment dzaima 2024-10-23 21:37:56 +0300
  • 7ac7297617 More simplifications to Singeli vector utilities Marshall Lochbaum 2024-10-21 21:43:59 -0400
  • 8099945068 Use vec_select from Singeli select includes Marshall Lochbaum 2024-10-21 15:35:03 -0400
  • 25724b6b3e More NEON functionality moved to neon_intrin/select Marshall Lochbaum 2024-10-13 21:25:22 -0400
  • fa43ffd882 Alias vec_shuffle to shuf Marshall Lochbaum 2024-10-12 20:48:33 -0400
  • 27af1c0558 Use Singeli select include for x86 blend, byteshift, zip Marshall Lochbaum 2024-10-12 15:41:48 -0400
  • 6140290813 Use Singeli select includes for shuffling Marshall Lochbaum 2024-10-12 12:47:51 -0400
  • 52f318e755 EACH_FILLS tests dzaima 2024-10-13 06:14:55 +0300
  • df1773712f
    Merge pull request #121 from svmhdvn/gmake-fix dzaima 2024-10-14 05:07:45 +0300
  • a18f547b44 build: fix on systems with bmake as make impl Siva Mahadevan 2024-10-13 21:46:49 -0400
  • 1e00aa71c9 atomEqualF & atomEEqual cleanup dzaima 2024-10-10 04:09:10 +0300
  • 503bddced5 bring back atomEqualF fast path dzaima 2024-10-10 03:59:02 +0300
  • 61e5ca2767 move function/modifier native ID into their respective structs dzaima 2024-10-10 03:46:12 +0300
  • 4885ca1c62 fix rtwrap decompose dzaima 2024-10-10 03:18:03 +0300
  • 82bf0d301c more comparison tests dzaima 2024-10-10 03:06:59 +0300
  • 5a6f3fd978 move comparison test section to beside total array comparison dzaima 2024-10-10 03:01:48 +0300
  • 4e071cf170 disable |, ⌊, ⌈, <, > wrapping for RT_VERIFY dzaima 2024-10-10 03:17:36 +0300
  • 6bd47b067e split apart def_decompose dzaima 2024-10-10 02:33:24 +0300