Commit Graph

  • 4bc5350747 Branchless sparse i16 Where Marshall Lochbaum 2022-09-11 11:00:17 -0400
  • 1f40d36712 Separate where into its own function and split different types completely Marshall Lochbaum 2022-09-11 08:31:20 -0400
  • 638121c054 Use blocked method for i32 Where with or without Singeli Marshall Lochbaum 2022-09-10 22:12:29 -0400
  • 4999758497 Branchless i32 sparse Where Marshall Lochbaum 2022-09-10 13:13:41 -0400
  • 017fcaea2a Move slash functions to a separate file Marshall Lochbaum 2022-09-10 09:58:42 -0400
  • bb3ecdca6a Sparse Where Marshall Lochbaum 2022-09-10 09:12:16 -0400
  • ae9817fd0d Fast path for / with empty result, and faster non-Singeli Where Marshall Lochbaum 2022-09-10 08:57:41 -0400
  • d8abc4ef7e Separate handling for ⊔ when 𝕨 changes value rarely Marshall Lochbaum 2022-09-09 18:11:12 -0400
  • ef2dee9a4d Special case data movement for ⊔ on sorted 𝕨 Marshall Lochbaum 2022-09-09 13:21:08 -0400
  • 569dcacac1 Do Group's range check and ¯1 counting on 1- or 2-byte ints Marshall Lochbaum 2022-09-08 21:52:34 -0400
  • 1e143644b6 Filter out ¯1s in Group with / if there are enough of them Marshall Lochbaum 2022-09-08 20:55:45 -0400
  • e266fbd09e basic float monadic arith loops, incl. -fno-math-errno dzaima 2022-09-08 23:18:47 +0300
  • 2358436b89 more efficient spreadBits{[8]u32, u8} dzaima 2022-09-08 21:51:25 +0300
  • d3a1fee437 singeli bit_sel dzaima 2022-09-08 21:39:51 +0300
  • d60713225b just use u64 in bitsel dzaima 2022-09-08 20:25:09 +0300
  • 1619b6087b mild optimization of v_set argument consumption dzaima 2022-09-08 20:05:13 +0300
  • 3ccef1fe86 ⊔𝕩 → ⊔⟜(↕≠)𝕩 dzaima 2022-09-08 14:52:19 +0300
  • 59cadb06fe fix ⟨⟨⟩⟩⊏⟨⟩ incorrectly erroring dzaima 2022-09-08 00:56:31 +0300
  • 66d32e0acb fast exit for depth-2 𝕨 in 𝕨⊏𝕩 dzaima 2022-09-08 00:47:25 +0300
  • 96ca9092ba attempt at better bit_sel dispatch dzaima 2022-09-08 00:37:44 +0300
  • 234443a17c src/README.md updates dzaima 2022-09-08 00:03:42 +0300
  • 59e947cc20 more consistent type checking & conversion functions dzaima 2022-09-07 23:56:03 +0300
  • 734e9fc116 better type checking helpers dzaima 2022-09-07 23:21:24 +0300
  • 2299c8094e make gcc scream less dzaima 2022-09-07 20:28:45 +0300
  • 0e306ca149 use WRAP instead of manual index management dzaima 2022-09-07 20:17:54 +0300
  • 10004b82f7 make 1.2‿3.4⊏≍˘"abcd" error dzaima 2022-09-07 19:53:00 +0300
  • 94e6168319 don't read past the end of array even if the value isn't used dzaima 2022-09-07 19:41:15 +0300
  • 5d660cbf19 deduplicate hashmap size doubling function dzaima 2022-09-07 19:27:14 +0300
  • 1a583725be use TALLOCP in other places dzaima 2022-09-07 17:59:17 +0300
  • 40b76b840e
    Merge pull request #41 from mlochbaum/select dzaima 2022-09-07 17:07:28 +0300
  • b784dac9e1 make clang not go crazy dzaima 2022-09-07 00:27:23 +0300
  • a43a42e4d1 Select from small boolean list by copying to bytes Marshall Lochbaum 2022-09-06 16:30:03 -0400
  • f877a45a63 Vectorize range checks for 1- and 2-byte indices Marshall Lochbaum 2022-09-06 16:06:00 -0400
  • 4a8244f2c1 Faster selection from boolean list Marshall Lochbaum 2022-09-06 12:42:19 -0400
  • 55263bcbc4 Handle bit selection from arrays with under 2 elements Marshall Lochbaum 2022-09-06 12:40:18 -0400
  • 1c279a0a20 fast bitarr×intarr dzaima 2022-09-06 22:47:16 +0300
  • a90e7de997 simpler atom-array arith on length 0 inputs dzaima 2022-09-06 18:22:58 +0300
  • 311aa74303 simpler bitsel for ∨ dzaima 2022-09-06 17:31:42 +0300
  • ce1c81219e manual wideness check dzaima 2022-09-06 17:23:57 +0300
  • 0b4259559c include errno.h in file.c unconditionally dzaima 2022-09-06 00:08:19 +0300
  • bffe16d6e9 fix overflowed & float functions dzaima 2022-09-05 23:23:51 +0300
  • c0b2164dc8 new array-atom arith dzaima 2022-09-05 22:52:53 +0300
  • 629076d001 new ⟨⟩↕0 behavior & fix /⁼large_i8arr dzaima 2022-09-05 22:40:29 +0300
  • 7f00d54372 rearrange arithd.c a bit more dzaima 2022-09-02 17:34:38 +0300
  • 47f48bbd94
    Merge pull request #40 from mlochbaum/ind-inv dzaima 2022-09-04 00:32:54 +0300
  • 7792c60180 Fix unnecessary errors with computed shape 0, such as 1e8‿⌊‿1e8⥊↕5 Marshall Lochbaum 2022-09-03 14:30:52 -0400
  • 97bfacec44 Fix hang in •rand.Subset caused by moving wi==0 case into wi==xi Marshall Lochbaum 2022-09-03 14:17:06 -0400
  • 44c0baee05 No need to initialize the negative half of the counts table Marshall Lochbaum 2022-09-03 14:07:17 -0400
  • 37af32d95c Faster /⁼ for boolean and long 1-byte and 2-byte 𝕩 Marshall Lochbaum 2022-09-02 11:27:53 -0400
  • 0139c5f58d separate dyadic arith cases dzaima 2022-09-02 14:01:35 +0300
  • a637adadd9 rearrange arithd.c in preparation for separating its cases dzaima 2022-09-02 13:10:04 +0300
  • 76141f6fe6 reduce lifetime of e in dyArith_AA dzaima 2022-09-01 22:45:30 +0300
  • 1b8bd0707e make dyarith array-array dispatch its own function dzaima 2022-09-01 22:22:13 +0300
  • bd04e38a5e use existing copy functions for arr_join_inline dzaima 2022-09-01 19:52:51 +0300
  • 90dd740bfc use cpyHArr instead of a getter loop dzaima 2022-09-01 19:24:34 +0300
  • dad47cf391 move arith_recd to a noinline function dzaima 2022-09-01 18:51:02 +0300
  • f758a3f1f6 proper argument name dzaima 2022-09-01 18:40:44 +0300
  • 1c43ad6431 squeeze •rand.Subset result dzaima 2022-09-01 18:25:10 +0300
  • b386531b30 don't re-link when not necessary dzaima 2022-09-01 18:11:24 +0300
  • d32e373168 attempt at fixing some makefile problems dzaima 2022-09-01 17:55:36 +0300
  • e72c7dc937 faster ÷ ⋆ √ | ⋆⁼ on i8 & i16 arrays dzaima 2022-09-01 15:59:26 +0300
  • 5cfa89ad5a separate functions for comparison scalar & array cases dzaima 2022-09-01 15:05:38 +0300
  • 72f25dad54 bitarr ↕1 and ↕2 dzaima 2022-09-01 13:29:03 +0300
  • 7d716d5ce0 elInt dzaima 2022-09-01 11:17:49 +0300
  • 203916b44f error on non-numbers in +𝕩 dzaima 2022-09-01 11:15:43 +0300
  • 6f707a5f9d
    Merge pull request #39 from mlochbaum/misc dzaima 2022-09-01 00:25:56 +0300
  • 76c93d85e1 Dyadic 32-bit / based on 8-bit monadic / Marshall Lochbaum 2022-08-31 15:12:05 -0400
  • 1e0cf78fd0 Implement 4-byte where with blocks of 2-byte where Marshall Lochbaum 2022-08-30 20:37:38 -0400
  • 6f02705d9f Make small-range Grade code style more like radix code Marshall Lochbaum 2022-08-30 18:50:21 -0400
  • 758c174cae Optimistic special case to quickly invert permutations with ⍋ Marshall Lochbaum 2022-08-30 18:39:36 -0400
  • a42335af4f Use or-equals for dyadic ∊ when 𝕩 is small and arguments are flat Marshall Lochbaum 2022-08-30 09:36:33 -0400
  • 15b444540e Extend self-search typed code to characters by treating as numbers Marshall Lochbaum 2022-08-29 21:05:46 -0400
  • 9028ebfe8b Native Windows (↕) Marshall Lochbaum 2022-08-29 07:38:36 -0400
  • 88f73d2585 fix comment & add note dzaima 2022-08-30 10:26:51 +0300
  • 9775339e2c squeeze result of ⌊ and ⌈ dzaima 2022-08-30 00:18:25 +0300
  • ce1032f80d second instruction shouldn't have a swap requirement anymore dzaima 2022-08-29 23:09:42 +0300
  • 8ea2352a7b clear out swap table after first instruction generation dzaima 2022-08-29 23:08:38 +0300
  • 70ee775a33 fix i16arr+i8arr overflow dzaima 2022-08-29 22:57:54 +0300
  • 39a0a894e2 correct char-char→signed int overflow checking dzaima 2022-08-29 22:48:46 +0300
  • c84d731017 disallow optimizing to fma dzaima 2022-08-29 22:47:21 +0300
  • b37f620889 disable manual array F character loops when not running singeli dzaima 2022-08-29 19:46:41 +0300
  • 78e5799147 merge dyarith.singeli and dyarith2.singeli dzaima 2022-08-29 19:16:35 +0300
  • 777c5cc1a5 rename things dzaima 2022-08-29 19:05:34 +0300
  • ce5ece34e1 new arith dzaima 2022-08-28 23:27:17 +0300
  • 8074a0c191
    Merge pull request #38 from mlochbaum/master dzaima 2022-08-28 16:50:15 +0300
  • 3c1ef1c9a2 use elNum dzaima 2022-08-28 16:44:21 +0300
  • f80d3ce263 Branchless integer binary search Marshall Lochbaum 2022-08-28 09:16:25 -0400
  • 139faee0a1 Unify integer and character Bins Marshall Lochbaum 2022-08-28 09:16:09 -0400
  • bcf3c4dbd0
    Merge pull request #37 from mlochbaum/sort dzaima 2022-08-28 00:06:24 +0300
  • 9376e02ecb Boolean sorting Marshall Lochbaum 2022-08-27 15:44:25 -0400
  • e027aca0cf Radix grades Marshall Lochbaum 2022-08-27 15:16:20 -0400
  • 36e54d1a39 1-byte radix with length ==256 doesn't work because of SWAR overflow when descending Marshall Lochbaum 2022-08-27 11:10:17 -0400
  • 4ec83e51a6 Merge sort.h into grade.h Marshall Lochbaum 2022-08-27 10:48:21 -0400
  • 9d59dff522 4-byte sorting: insertion and radix Marshall Lochbaum 2022-08-26 20:54:12 -0400
  • f946a3de41 2-byte sorting: insertion, radix, counting Marshall Lochbaum 2022-08-26 20:22:08 -0400
  • 5034e55c79 Move Sort to template file sort.h (for up/down) like Grade Marshall Lochbaum 2022-08-26 17:25:40 -0400
  • a0a4ecadc7 1-byte sorting: insertion, bucket, counting Marshall Lochbaum 2022-08-26 16:07:00 -0400
  • 5a162c7929 macos-specific sub-second file times dzaima 2022-08-26 22:46:00 +0300
  • 2939395906 fallback file time getting method dzaima 2022-08-26 22:29:45 +0300
  • 81e09956d3 •file.Size, make •file.List decode UTF-8 dzaima 2022-08-26 13:37:51 +0300