Commit Graph

  • fce7567349 widen bits in bits⊸⊏˘mat to i8 dzaima 2024-07-25 05:32:35 +0300
  • 90646563a0 don't use zip method for select-cells dzaima 2024-07-24 23:32:54 +0300
  • 4a874710c7 allow unhandled types for select_rows dzaima 2024-07-23 19:01:35 +0300
  • 478c389c4b fast inds⊸⊏˘ mat dzaima 2024-07-25 17:34:15 +0300
  • 0d7bf86182 baseline inds⊸⊏˘ mat dzaima 2024-07-26 22:15:59 +0300
  • f7dd900b3a faster 𝕨⊏𝕩, incl. aarch64 support dzaima 2024-07-25 17:33:44 +0300
  • afa3353b35 fix makefile having incorrect name of stored-warn dzaima 2024-07-26 00:00:42 +0300
  • cda5fd8595 make native compiler not error in debug builds dzaima 2024-07-23 20:29:28 +0300
  • 3689e6105c support 2-modifiers in native compiler dzaima 2024-07-23 19:57:24 +0300
  • b7f7eda1ff more cells.h comments dzaima 2024-07-25 16:45:29 +0300
  • d6483a93d3 x86 unzip & unzip128 dzaima 2024-07-23 17:20:51 +0300
  • 2ed06e0db6 get_range Singeli binding dzaima 2024-07-23 17:19:07 +0300
  • 7bb5a1d9d8 more tests for ⊏ dzaima 2024-07-25 23:55:54 +0300
  • 2ca488dd66 aarch64 bool128 select dzaima 2024-07-18 21:44:34 +0300
  • 0d5e77766a build select.singeli everywhere dzaima 2024-07-23 16:48:12 +0300
  • f1f4f1734e extract m_tyslice dzaima 2024-07-23 15:49:36 +0300
  • b99d5f23d5 extract bitsel to a function table dzaima 2024-07-23 15:38:36 +0300
  • 2b323817ca make 𝕨⊏𝕩 error messages not differ across types dzaima 2024-07-23 17:42:55 +0300
  • e757b690dd mention src/singeli/ in src/README.md dzaima 2024-07-20 22:34:18 +0300
  • c274c35b1e Singeli @for_blocks dzaima 2024-07-23 17:23:22 +0300
  • 079d3ffe91 better aarch64 homMask{[16]u8,[16]u8} dzaima 2024-07-18 23:54:36 +0300
  • 0b69fa84cd NEON comparisons to 0 dzaima 2024-07-18 21:56:24 +0300
  • c0573d602c don't error on non-vector types checked by genchk dzaima 2024-07-20 21:53:46 +0300
  • 6fcf9d1fd1 update Singeli submodule dzaima 2024-07-25 16:35:09 +0300
  • a9fbd0301d fix ∾˝⎉k refcounting dzaima 2024-07-15 00:41:40 +0300
  • a9f59b5793 mention target_os in README dzaima 2024-07-15 00:17:40 +0300
  • 3e9af0a910 fast-path for n•rand.Range 1 dzaima 2024-07-13 19:57:27 +0300
  • 2dd3798228 more Singeli cleanup dzaima 2024-07-15 00:54:49 +0300
  • 2436c45f1d Singeli ceil_log2 for compile-time numbers dzaima 2024-07-13 00:07:59 +0300
  • 66a8529d88 remove redundant parentheses in Singeli destructuring dzaima 2024-07-13 00:07:42 +0300
  • 59515acc90 test variations of erroring ⊏ dzaima 2024-07-13 00:00:49 +0300
  • ed2198c06a disable tail calls in bqn_fmt & bqn_repr dzaima 2024-07-12 18:07:00 +0300
  • f79e4c6d0a make /(m?zip(128)?|pack)(Lo|Hi)/ not have Lo/Hi as part of name dzaima 2024-07-12 05:23:46 +0300
  • a671ec9462 aarch64 bittab dzaima 2024-06-16 13:58:50 +0300
  • 6dbe76727f update offset for stacktraces on FLDG dzaima 2024-07-10 20:22:00 +0300
  • 838639e610 fix gcc-9 build dzaima 2024-07-10 03:51:20 +0300
  • 7aab7ceb8b cleanup dzaima 2024-07-10 03:44:02 +0300
  • 2bc1ba9f4a fix aarch64 squeeze dzaima 2024-07-10 03:28:57 +0300
  • 52d1e61fe0 don't bother early-exiting on masked iteration dzaima 2024-07-10 03:24:10 +0300
  • 2fdd72f1c4 faster <SSE4.1 integer accumulating in squeeze dzaima 2024-07-10 02:40:54 +0300
  • a45925b0c6 fix SSE2 squeeze dzaima 2024-07-10 02:10:55 +0300
  • f941245b34 avoid needing two final folds on x86 dzaima 2024-07-10 01:55:23 +0300
  • 40a5c40bda faster squeeze dzaima 2024-07-10 03:17:20 +0300
  • d66be091be shufps Singeli mapping dzaima 2024-07-08 22:49:01 +0300
  • e96948cb64 Singeli zip128 dzaima 2024-07-05 21:42:21 +0300
  • a8af920dc9 •internal.PureKeep dzaima 2024-07-09 02:00:19 +0300
  • 01b1ee3637 make ⊢¨ more consistent with other PureIdentFn¨ dzaima 2024-07-09 01:59:19 +0300
  • 826179c83d •internal utility for getting & setting object flags dzaima 2024-07-08 22:52:15 +0300
  • 352f22f8f6 don't vectorize tiny loop dzaima 2024-07-04 21:52:56 +0300
  • 3e24442011 use ux in base cmp dzaima 2024-07-08 04:25:15 +0300
  • c466db24bc
    Merge pull request #114 from mlochbaum/simplify dzaima 2024-07-08 04:23:42 +0300
  • d625fc0d21 Some tests for rank, type, etc. errors in cells/rank Marshall Lochbaum 2024-07-07 15:52:33 -0400
  • 201fe1e57c Extend fold_rows to any rank, as long as stride is 1 Marshall Lochbaum 2024-07-04 15:23:41 -0400
  • 48d77e722f Extend ˝ to ´ conversion (for singleton cells) to any rank Marshall Lochbaum 2024-07-04 12:03:26 -0400
  • 46704defb6 Convert list⊸↑˘ to a single ↑ call, likewise ↓ and ⌽ Marshall Lochbaum 2024-07-04 11:40:05 -0400
  • ce98efc525 Transform atom⊸⌽⎉k into a single high-rank ⌽ call Marshall Lochbaum 2024-07-04 09:58:14 -0400
  • 3b49294145 Specialize ApdMut to reshape if there's only one append Marshall Lochbaum 2024-07-03 22:17:06 -0400
  • 76a40491e9 Remove some rank restrictions from for_cells_SA Marshall Lochbaum 2024-06-26 19:50:56 -0400
  • 4450d36187 Better handling for «˘ and »˘ trivial cases Marshall Lochbaum 2024-06-26 14:59:45 -0400
  • d89a6da2da Evaluate monadic «» on a length-1 argument by reshaping the fill Marshall Lochbaum 2024-06-26 14:42:57 -0400
  • 6eca24ce0a Special code for ∾˝˘ and fold-cells of empty Marshall Lochbaum 2024-06-26 08:40:49 -0400
  • 9d5caad9a9 Compute Insert identity natively Marshall Lochbaum 2024-06-23 10:51:01 -0400
  • 5d4391a3fd Combine monadic ˘ with outer ˘ or ⎉ Marshall Lochbaum 2024-06-23 10:10:00 -0400
  • a19b5401f7 Fast cases for monadic rank with cell length ≤1 Marshall Lochbaum 2024-06-22 14:55:07 -0400
  • 8766810db8 Rank-agnostic select_cells Marshall Lochbaum 2024-06-22 09:39:02 -0400
  • ab15eaeeeb Length-1 reverse is a no-op Marshall Lochbaum 2024-06-22 09:38:38 -0400
  • 2ce983041c fix shared •HashMap value list dzaima 2024-06-30 14:06:05 +0300
  • 9c05b68870 ubsan option for build.bqn dzaima 2024-06-29 17:12:12 +0300
  • 47b5d7d832 fix harmless copy-paste error dzaima 2024-06-29 04:38:44 +0300
  • d8a79145a6 add no-catch option to test/run.bqn dzaima 2024-06-29 02:59:44 +0300
  • 0319be27ea fix bad error message on /⁼3‿3⥊0 dzaima 2024-06-29 02:59:15 +0300
  • 7eb60d4ec1 correct ElType comment dzaima 2023-12-11 14:20:09 +0200
  • d885a7d965 don't leave the temporary bootstrap CBQN around dzaima 2024-06-26 06:41:27 +0300
  • 79127db4a2
    Merge pull request #113 from mlochbaum/fold-rows dzaima 2024-06-26 06:41:00 +0300
  • 6824b9a747 fix another unaligned load dzaima 2024-06-26 06:09:30 +0300
  • 7474e51642 rename mask_of in spaced.singeli dzaima 2024-06-26 05:23:41 +0300
  • 452065adaf add magic incantation to make optimizer happy dzaima 2024-06-26 05:00:11 +0300
  • e365ecf7dc Fix F`˘ on rank 0 Marshall Lochbaum 2024-06-25 22:31:47 -0400
  • ace6105bf6 Fixes for ∧˝˘ and ∨˝˘ on long rows Marshall Lochbaum 2024-06-25 22:30:56 -0400
  • 445da5a659 actually test for correctness dzaima 2024-06-26 04:21:46 +0300
  • 8e7f19a883 make cells test work under heapverify dzaima 2024-06-26 03:43:10 +0300
  • b6bf3f2da6 suppress clang ubsan warning dzaima 2024-06-26 03:35:25 +0300
  • 8d5efa72f0 testing of ˝˘ & `˘ dzaima 2024-06-26 03:16:41 +0300
  • 024ab669d5 fix ∧˝˘0‿8⥊0 running comparison functions on length-0 inputs dzaima 2024-06-26 03:08:58 +0300
  • 6c110a29ee use loadu_* & storeu_* for unaligned loads/stores dzaima 2024-06-26 03:07:02 +0300
  • ace06e96ae document CMP_* non-zero length requirement dzaima 2024-06-26 03:09:30 +0300
  • 7e15dfe104 make loadu_*/storeu_* take void* dzaima 2024-06-26 03:03:57 +0300
  • e952e4ca3e Disable F´˘ special code when cell rank isn't 1 Marshall Lochbaum 2024-06-23 17:06:37 -0400
  • 5758853ccb Can't call fold-rows special code with reduction length 0 Marshall Lochbaum 2024-06-22 21:57:59 -0400
  • 5f1c677d48 Support ∧∨ synonyms in fold_rows_bit Marshall Lochbaum 2024-06-20 11:00:57 -0400
  • 355efd1eb4 Insert-cells and scan implementation comments, and minor tweaks Marshall Lochbaum 2024-06-19 18:36:53 -0400
  • 4b0f105a7f Implement short-row num⊏˘bool (including ⊣˝˘ ⊢˝˘) with fold code Marshall Lochbaum 2024-06-18 14:05:16 -0400
  • cb1b72fbb2 Extend boolean F˝˘ special code to any ranks Marshall Lochbaum 2024-06-18 07:46:18 -0400
  • 1e6c7057e8 Extend boolean F`˘ special code to any frame and cell rank Marshall Lochbaum 2024-06-17 21:51:32 -0400
  • 5e3cc6de81 Refactoring, simplification, comments Marshall Lochbaum 2024-06-17 11:29:40 -0400
  • a8b036ad08 Implement ≠˝˘ and =˝˘ like ∧∨ on width<64 boolean Marshall Lochbaum 2024-06-17 09:18:06 -0400
  • f0f130c42e Use a lookup table for evenly-spaced masks, getting division results from those Marshall Lochbaum 2024-06-15 18:30:46 -0400
  • e6f1e04de2 Fast generic and pext-based ∧˝˘ and ∨˝˘ on width<64 Marshall Lochbaum 2024-06-13 16:52:52 -0400
  • 40bf3bfd1c AVX2-based boolean short-row scans Marshall Lochbaum 2024-06-04 20:35:46 -0400
  • 6d27dd391b Clean up scan_rows_bit logic and implement -˘ as (2×⊣˘)-+`˘ Marshall Lochbaum 2024-06-03 18:49:32 -0400