Commit Graph

  • b4fc29e295 don't use <esc>[2K for clearing lines dzaima 2022-12-26 17:56:29 +0200
  • 7bcee07489 initialize shape properly in one case of high-rank take/drop dzaima 2022-12-26 16:02:13 +0200
  • f6b00f97ee make git invocation in build.bqn working-directory-independent dzaima 2022-12-25 13:44:09 +0200
  • 326d54e130 native list↑𝕩 & list↓𝕩 dzaima 2022-12-25 13:38:43 +0200
  • 8fad76762a reorder take_c2 dzaima 2022-12-24 17:53:16 +0200
  • e2b45e33e1 simplify splitCells dzaima 2022-12-24 17:29:48 +0200
  • 424448aa40 remove rt_memberOf & rt_indexOf dzaima 2022-12-24 17:12:35 +0200
  • 95b1beb9db native high-rank ∊ & ⊐ dzaima 2022-12-24 17:07:25 +0200
  • 3f40955d83 reorder indexOf_c2 dzaima 2022-12-24 17:04:57 +0200
  • e3b5af8a44 fix 𝕨∊p & p⊐𝕩 with an empty high-rank p dzaima 2022-12-24 16:54:04 +0200
  • 24dba9ba40 fix a bunch of small memory leaks around REPLXX dzaima 2022-12-24 16:42:42 +0200
  • e2c21f6103 make build/build have exit code 1 on failed build dzaima 2022-12-24 16:16:31 +0200
  • c58ba6ee52 move global build time start to earliest possible point dzaima 2022-12-24 15:52:18 +0200
  • 4b1b935742 remove comment about PIE= dzaima 2022-12-24 15:47:12 +0200
  • 11005b8bf8 remove default -no-pie dzaima 2022-12-24 15:46:34 +0200
  • 3d34155f83 actually add -O2 to make forbuild build dzaima 2022-12-24 15:28:55 +0200
  • bb7fe02a45 separate checks for x86-64 & NEON Singeli importing dzaima 2022-12-24 15:27:53 +0200
  • 6d79ce9fb1 use arch-independent names for things suppported on both x86-64 & NEON dzaima 2022-12-24 15:26:05 +0200
  • d38316e670 BQN-based build system dzaima 2022-12-23 21:10:52 +0200
  • efcc6a28be add •file.path to autocomplete options dzaima 2022-12-23 18:08:33 +0200
  • bbc66cbf3f fix display of running multiline code while cursor isn't in last line dzaima 2022-12-23 18:06:51 +0200
  • 381ce70212 backslash+enter → literal newline dzaima 2022-12-23 18:02:18 +0200
  • baad87c706 clean up make clean= & builddir= dzaima 2022-12-20 19:52:41 +0200
  • 2d1882fae2 more easily configurable job count dzaima 2022-12-20 19:08:48 +0200
  • a6b22c75fd fast list⊐unit for more list types dzaima 2022-12-20 18:32:51 +0200
  • e69d8dc6e9 native ⌾(v⊸⌽) dzaima 2022-12-20 18:31:34 +0200
  • d05621f8f4 generic vector width calculation for NEON-supported files dzaima 2022-12-19 23:56:07 +0200
  • b91e72567d update Singeli submodule, conditional NEON includes dzaima 2022-12-19 23:31:22 +0200
  • 43be832a37 fix highrank∊list deshaping the result in some scenarios dzaima 2022-12-18 23:58:10 +0200
  • 8188886538 fast 𝔽´⟨x⟩ & 𝔽´⟨x,y⟩ dzaima 2022-12-18 23:49:09 +0200
  • aa4b92cb01 don't define •SH if not available dzaima 2022-12-18 23:40:55 +0200
  • 29cecb0b6e remove •PrimInd dzaima 2022-12-18 23:16:36 +0200
  • 6380364e62 native highrank∊vector dzaima 2022-12-18 23:05:58 +0200
  • 30a6de0a31 add unrolling to aarch64 cmp.singeli dzaima 2022-12-18 22:30:31 +0200
  • f9c919d5ce getmask → homMask/topMask dzaima 2022-12-18 22:27:42 +0200
  • a07b94b889 more general equal.singeli, remove trailing semicolons dzaima 2022-12-18 21:53:25 +0200
  • 2a32f52f8d blend renaming dzaima 2022-12-18 02:49:42 +0200
  • 09b3279958 unified Singeli including mechanism dzaima 2022-12-18 01:51:06 +0200
  • 9768526ddc move last bit of equal.singeli to loadBatch dzaima 2022-12-18 01:28:27 +0200
  • 40f36fa366 Singeli-generated table for bits.singeli dzaima 2022-12-17 17:51:56 +0200
  • 2c73199ed4 use Singeli-generated function tables for comparison dzaima 2022-12-16 18:51:48 +0200
  • f49dd5e8cf move cmp.singeli NaN check outside for non-x86_64 dzaima 2022-12-16 16:33:37 +0200
  • de1bc73320 update Singeli submodule dzaima 2022-12-16 16:30:59 +0200
  • 6e3cb9a425 andIsZero→andAllZero; NEON andnz, blendF, sel, bitops.singeli dzaima 2022-12-16 00:44:32 +0200
  • cbd891b768 NEON widen, narrow, undefPromote; make undefPromote more restrictive dzaima 2022-12-15 23:19:00 +0200
  • 6fac426f90 add v_dbl & v_half, NEON half{x,n} dzaima 2022-12-15 21:57:34 +0200
  • 791e7d122f ucvt→narrow dzaima 2022-12-15 21:44:48 +0200
  • 2b7291dbfb cvt{E,T,v} → widen{T,to_el{E,v}} dzaima 2022-12-15 20:26:41 +0200
  • 223bd0627a update Singeli submodule dzaima 2022-12-15 16:15:50 +0200
  • 9a9fade18a NEON loadLow, storeLow, andIsZero, maskstore, pack dzaima 2022-12-15 16:03:53 +0200
  • 6f2d6bfb0b remove offset argument from maskstore; don't use w256 in mask.singeli dzaima 2022-12-15 15:48:16 +0200
  • ddf1dca0b4 update Singeli submodle dzaima 2022-12-15 02:16:45 +0200
  • 77edc734e8 move storeLow & loadLow to sse3 & avx; fix NEON __not dzaima 2022-12-15 00:37:17 +0200
  • 9fea5f1e13 neg(Any|All) → top(Any|All), msk(Any|All) → hom(Any|All) dzaima 2022-12-13 21:46:50 +0200
  • 336da996b3 any/all vector check renaming dzaima 2022-12-13 02:03:11 +0200
  • 7e13bf7f45 more neon.singeli dzaima 2022-12-12 02:59:11 +0200
  • 444268f051 width{eltype{T}} → elwidth{T} dzaima 2022-12-12 00:07:50 +0200
  • 2a1611ac81 start of neon.singeli dzaima 2022-12-11 23:58:35 +0200
  • 982a54d98a update bytecode dzaima 2022-12-11 15:36:26 +0200
  • 0ef8800530 native F⌾⊢𝕩 & F⌾⊣𝕩 dzaima 2022-12-08 13:26:04 +0200
  • 3a3f213a92 assign broadcast{T,v} to T**v dzaima 2022-12-07 23:53:19 +0200
  • ce45d487a3 use mutating assignment dzaima 2022-12-07 23:24:26 +0200
  • 07d4567a52 use spread arguments dzaima 2022-12-07 23:10:17 +0200
  • bf71fdb4f8 singeli now formats generators itself dzaima 2022-12-07 22:41:17 +0200
  • 1b88c4cc62 don't allocate twice the memory requested dzaima 2022-12-07 22:22:34 +0200
  • 49c0d9a355 update replxx submodule dzaima 2022-12-05 15:30:15 +0200
  • af978bbc91 cave in and initialize variables in gcc dzaima 2022-12-03 23:58:33 +0200
  • 6ca5050f4d reduce gcc warnings dzaima 2022-12-03 23:43:18 +0200
  • e4f4864ed2 define CBQN_EXPORT by default, add macro test for exactly gcc dzaima 2022-12-03 23:19:24 +0200
  • 63e5e843ba fix f64arr/𝕩 dzaima 2022-12-03 23:14:13 +0200
  • 5d61785912 some more x86_64.h reordering dzaima 2022-12-03 22:54:49 +0200
  • 21984e1700 more x86_64.h cleanup dzaima 2022-12-03 15:21:05 +0200
  • 95f9e781fa fix )profiler output merging all REPL entries dzaima 2022-12-03 14:14:32 +0200
  • 732d1fdb34 remove ASMI/macro indirection dzaima 2022-12-02 17:25:47 +0200
  • bca0efb46c clean up x86_64.h dzaima 2022-12-02 17:10:04 +0200
  • 0023e46ab8 config file & light theme dzaima 2022-11-27 17:37:24 +0200
  • ae6e3a6504
    Merge pull request #64 from mlochbaum/avx2sel dzaima 2022-12-02 16:42:12 +0200
  • 7c6676a492 copy over custom _mm_loadu_si32 for old gcc dzaima 2022-12-01 13:37:01 +0200
  • 4077596b89 Sparse initialization for ⌾(i⊸⊏) byte table Marshall Lochbaum 2022-11-30 16:31:32 -0500
  • d7a06befb5 Select implementation comments Marshall Lochbaum 2022-11-30 16:15:50 -0500
  • 534c92f38a Select from <=128 booleans with shuffles Marshall Lochbaum 2022-11-29 22:03:37 -0500
  • fdfeb67e8f 2-register permutevar8x32 Marshall Lochbaum 2022-11-29 20:27:45 -0500
  • 41f464795f Selection with permutevar8x32 Marshall Lochbaum 2022-11-29 17:48:24 -0500
  • c17448ed71 Faster bit selection Marshall Lochbaum 2022-11-29 11:29:25 -0500
  • dc6e49b289 Shuffle selection using >1 data register Marshall Lochbaum 2022-11-28 21:38:36 -0500
  • b651ca73d3 Move base{} utility to base.singeli Marshall Lochbaum 2022-11-28 12:14:30 -0500
  • 43fba7cd78 Extend shuffle selection to >1 byte data Marshall Lochbaum 2022-11-28 11:28:29 -0500
  • 6639c17b81 1-byte select with <=16 elements using shuffle Marshall Lochbaum 2022-11-27 21:03:58 -0500
  • e09d555eba Refactoring and renaming in select.singeli Marshall Lochbaum 2022-11-26 20:39:18 -0500
  • 7cb8cf5772 Enable selection with AVX2 gather for i8 to i32 indices and 1- to 8-byte values Marshall Lochbaum 2022-11-26 20:30:11 -0500
  • dbc7c83f70 move usum to fold.c dzaima 2022-11-25 20:03:41 +0200
  • afe6c0eba9 use 1•internal.Info for rtverify fails dzaima 2022-11-25 19:53:14 +0200
  • c0e06833c0 move B squeeze non-integer case out of twice-generated generator dzaima 2022-11-25 19:38:38 +0200
  • b0fcc017ce clean up squeeze.singeli dzaima 2022-11-25 19:32:46 +0200
  • 0f3f91105e use anynePositive for squeeze.singeli dzaima 2022-11-25 19:12:24 +0200
  • 15f6dd5d48 simplify 16-bit anynePositive dzaima 2022-11-25 18:57:03 +0200
  • 5f85305f43 GPR mask testing dzaima 2022-11-25 02:20:58 +0200
  • efd79881cb use single mask array for all type widths dzaima 2022-11-25 02:16:33 +0200
  • cc0aa486db
    Merge pull request #63 from mlochbaum/fact-inv dzaima 2022-11-27 15:45:30 +0200
  • d911eafd31 Inverses for •math.Fact and •math.LogFact Marshall Lochbaum 2022-11-26 14:17:27 -0500