Commit Graph

  • af61c9bae3 working on target-nvptx64... need to add nvptx64 egaburov 2013-07-28 15:50:08 +02:00
  • 67b549a937 Added nvptx64 target. Things to do: egaburov 2013-07-28 14:31:43 +02:00
  • b6df447b55 Add reduce_add() for int8 and int16 types. Matt Pharr 2013-07-25 09:11:39 -07:00
  • 2d063925a1 Explicitly call the PBLENDVB intrinsic for i8 blending with sse4-8. Matt Pharr 2013-07-24 15:10:08 -07:00
  • bba84f247c Improved optimization of vector select instructions. Matt Pharr 2013-07-24 15:08:07 -07:00
  • 780b0dfe47 Add SSE4-16 target. Matt Pharr 2013-07-24 09:40:50 -07:00
  • 04d61afa23 Fix bug in lEmitVaryingSelect() for targets with i1 mask types. Matt Pharr 2013-07-25 09:40:48 -07:00
  • 663ebf7857 Merge pull request #551 from mmp/constfold Dmitry Babokin 2013-07-24 10:27:04 -07:00
  • 53414f12e6 Add SSE4 target optimized for computation with 8-bit datatypes. Matt Pharr 2013-07-23 17:30:32 -07:00
  • 15a3ef370a Use @llvm.readcyclecounter to implement stdlib clock() function. Matt Pharr 2013-07-23 17:11:01 -07:00
  • c14659c675 Fix bug in lGetConstantInt() in parse.yy. Matt Pharr 2013-07-23 17:02:49 -07:00
  • f7f281a256 Choose type for integer literals to match the target mask size (if possible). Matt Pharr 2013-07-23 17:01:03 -07:00
  • 9ba49eabb2 Reduce estimated costs for 8 and 16-bit min() and max() in stdlib. Matt Pharr 2013-07-23 16:52:43 -07:00
  • e7abf3f2ea Add support for mask vectors of 8 and 16-bit element types. Matt Pharr 2013-07-23 16:38:10 -07:00
  • 83e1630fbc Add support for fast division of varying int values by small constants. Matt Pharr 2013-07-23 16:49:56 -07:00
  • 0277ba1aaa Improve warnings for right shift by varying amounts. Matt Pharr 2013-07-23 16:49:07 -07:00
  • 753c001e69 Merge branch 'master' of https://github.com/ispc/ispc into constfold Matt Pharr 2013-07-23 16:12:04 -07:00
  • 10c0b42d0d Merge pull request #549 from mmp/fix-tot Dmitry Babokin 2013-07-23 09:14:08 -07:00
  • 564e61c828 Improvements to constant folding. Matt Pharr 2013-07-22 16:12:02 -07:00
  • 946c39a5df Fix build with LLVM top-of-tree. Matt Pharr 2013-07-22 15:42:52 -07:00
  • 2948e84846 Merge pull request #547 from mmp/arm-merge Jean-Luc Duprat 2013-07-22 09:24:16 -07:00
  • 068fd8098c Explicitly set armv7-eabi target triple on ARM. Matt Pharr 2013-07-20 11:19:10 -07:00
  • d7b0c5794e Add support for ARM NEON targets. Matt Pharr 2013-07-19 11:06:11 -07:00
  • b007bba59f Replace inline assembly in task system with equivalent gcc intrinsics. Matt Pharr 2013-07-19 10:08:28 -07:00
  • abf43ad01d Merge pull request #546 from dbabokin/release Dmitry Babokin 2013-07-19 18:49:07 -07:00
  • 922895de69 Changing ISPC version to 1.4.5dev Dmitry Babokin 2013-07-19 18:47:43 -07:00
  • 28f0bce9f2 Release 1.4.4 v1.4.4 Dmitry Babokin 2013-07-19 16:22:10 -07:00
  • 0f82f216a2 Merge pull request #544 from mmp/master Dmitry Babokin 2013-07-18 11:46:11 -07:00
  • 7454b1399c Handle SHL with a constant vector in LLVMVectorIsLinear(). Matt Pharr 2013-07-17 14:12:43 -07:00
  • 4ebf46bd63 Merge pull request #543 from mmp/master jbrodman 2013-07-17 10:38:06 -07:00
  • f1cce0ef5f Fix build with LLVM top-of-tree Matt Pharr 2013-07-17 09:25:00 -07:00
  • 8c9e873c10 Merge pull request #540 from dbabokin/embree_bug Dmitry Babokin 2013-07-04 10:45:06 -07:00
  • c85439e7bb Fix for the bug introduced by --intrumentation fix Dmitry Babokin 2013-07-04 21:41:57 +04:00
  • fd7f87b55e Supporting perf.py on Windows and some small corrections in it Ilia Filippov 2013-07-02 19:23:18 +04:00
  • 8be4128c5a Merge pull request #534 from ifilippov/perf Dmitry Babokin 2013-07-01 05:09:03 -07:00
  • 806e37338c add script for measuring performance Ilia Filippov 2013-07-01 13:30:49 +04:00
  • ec1095624a Merge pull request #527 from tkoziara/master Dmitry Babokin 2013-06-25 10:11:39 -07:00
  • a23d69ebe8 Copyright changed to simplify legal matters. Tomasz Koziara 2013-06-25 17:28:27 +01:00
  • 0aff61ffc6 Merge pull request #533 from dbabokin/patch Dmitry Babokin 2013-06-25 08:50:32 -07:00
  • 05aa540984 Quick fix for LLVM 3.3 patch Dmitry Babokin 2013-06-25 19:49:41 +04:00
  • 033e83e490 Merge pull request #532 from dbabokin/release_1_4_3 Dmitry Babokin 2013-06-25 07:42:08 -07:00
  • 594485c38c Release 1.4.3 v1.4.3 Dmitry Babokin 2013-06-25 18:24:44 +04:00
  • d52e2d5a8d License update (just dates) Dmitry Babokin 2013-06-25 17:02:42 +04:00
  • 1e5d852e2f Merge pull request #531 from ifilippov/qsize_fail Dmitry Babokin 2013-06-25 05:36:45 -07:00
  • cc32d913a0 replacement of qsize due to it's fails on MacOS Ilia Filippov 2013-06-25 16:27:25 +04:00
  • fc66066d4d Merge pull request #530 from dbabokin/llvm_fix Dmitry Babokin 2013-06-25 05:22:09 -07:00
  • 6169338815 Adding LLVM patch to fix #519 with LLVM 3.3 Dmitry Babokin 2013-06-25 16:21:14 +04:00
  • 86ee8db778 Parallel prefix sum added + minor amendements. Tomasz Koziara 2013-06-25 12:45:51 +01:00
  • 6bc8cb1ff1 Merge pull request #529 from ifilippov/instrument_fix Dmitry Babokin 2013-06-25 03:08:02 -07:00
  • 0fc49b1c37 Merge pull request #528 from ifilippov/test3 Dmitry Babokin 2013-06-25 02:14:24 -07:00
  • 9fb981e9a0 correction of --instrument option support Ilia Filippov 2013-06-25 11:51:35 +04:00
  • cba1b3cedd additional libraries for LLVM_3_4 build Ilia Filippov 2013-06-19 16:03:26 +04:00
  • 12c4512932 adding two additional libraries for LLVM_3_4 build Ilia Filippov 2013-06-19 14:15:58 +04:00
  • f2452f040d First commit of the radix sort example. Tomasz Koziara 2013-06-24 18:37:44 +01:00
  • 0dd1dbb568 Merge pull request #526 from dbabokin/master Dmitry Babokin 2013-06-23 23:10:19 -07:00
  • fdcec5a219 Tracking LLVM trunk: removing llvm::createSimplifyLibCallsPass() call Dmitry Babokin 2013-06-24 10:08:06 +04:00
  • bebab7ab0d Merge pull request #525 from dbabokin/debug Dmitry Babokin 2013-06-21 03:56:17 -07:00
  • fb771b6aa3 --debug output: stdout instead of stderr Dmitry Babokin 2013-06-20 22:47:29 +04:00
  • 8156559475 Merge pull request #522 from dbabokin/broadcast jbrodman 2013-06-18 11:47:24 -07:00
  • 9f5e51cd01 Merge pull request #523 from dbabokin/tot jbrodman 2013-06-18 11:47:16 -07:00
  • 27daab2f1b Fix for #520 Dmitry Babokin 2013-06-18 22:15:49 +04:00
  • c4d404b15f Tracking ToT changes: changes in MCContext interface Dmitry Babokin 2013-06-18 22:13:14 +04:00
  • 95fcdc36ee Tracking ToT changes, which now require to link option library. This is Unix only. Windows will be fixed separately Dmitry Babokin 2013-06-18 22:12:33 +04:00
  • 2fdaba53c1 Merge pull request #517 from ifilippov/bug_34 Dmitry Babokin 2013-06-14 08:40:01 -07:00
  • 5c89080469 changes in function LLVMFlattenInsertChain Ilia Filippov 2013-06-14 16:38:54 +04:00
  • d92f9df17c changes in function LLVMFlattenInsertChain Ilia Filippov 2013-06-14 15:21:45 +04:00
  • f551390420 Merge pull request #516 from ifilippov/master Dmitry Babokin 2013-06-13 08:48:29 -07:00
  • 8642b4d89f changing run_tests to support skipping tests and time Ilia Filippov 2013-06-13 19:25:34 +04:00
  • 6fb70c307d changing run_tests to support skipping tests and time Ilia Filippov 2013-06-13 19:00:02 +04:00
  • d08346fbcf changes to support skipping tests Ilia Filippov 2013-06-13 16:47:10 +04:00
  • 141d240a91 Merge pull request #513 from dbabokin/release_142 jbrodman 2013-06-11 07:47:37 -07:00
  • cf9ceb6bf9 Release 1.4.2, 11 June 2013 v1.4.2 Dmitry Babokin 2013-06-11 17:18:54 +04:00
  • 7589ae0de5 Merge pull request #512 from ifilippov/bug_34 Dmitry Babokin 2013-06-04 07:10:04 -07:00
  • f46e5b37e9 Merge pull request #511 from dbabokin/win32 jbrodman 2013-06-04 06:43:53 -07:00
  • 560acd5017 changes to support createFunction() with DICompositeType argument in LLVM_3_4 Ilia Filippov 2013-06-04 15:48:39 +04:00
  • 2267f278d2 Fix for #503 - avoid omitting frame pointer on Win32 Dmitry Babokin 2013-06-04 14:38:56 +04:00
  • 0feeef585c Merge pull request #509 from jbrodman/master jbrodman 2013-05-30 13:21:23 -07:00
  • 6211966c55 Change mask to use __mmask16 instead of a struct. james.brodman 2013-05-30 16:04:44 -04:00
  • 92f591b4bd Merge pull request #508 from dbabokin/master Dmitry Babokin 2013-05-28 08:59:13 -07:00
  • 29ceb42b7b Bumping version to 1.4.1dev Dmitry Babokin 2013-05-28 19:58:27 +04:00
  • adaabe5993 Merge pull request #507 from dbabokin/master Dmitry Babokin 2013-05-28 08:49:14 -07:00
  • 6c392ee4a1 Changes for 1.4.1 release v1.4.1 Dmitry Babokin 2013-05-28 19:46:30 +04:00
  • 7699eda5ba Merge pull request #506 from jbrodman/master jbrodman 2013-05-28 08:13:03 -07:00
  • d8b5fd5409 Typo fix. james.brodman 2013-05-28 11:13:43 -04:00
  • b37ffdbe85 Merge pull request #505 from dbabokin/release Dmitry Babokin 2013-05-27 06:03:22 -07:00
  • 481bcc732b Changes for 1.4.0 release v1.4.0 Dmitry Babokin 2013-05-27 16:48:41 +04:00
  • ce175aee4c Merge pull request #504 from dbabokin/attr jbrodman 2013-05-24 07:38:03 -07:00
  • 50896b373b Merge pull request #502 from dbabokin/malloc jbrodman 2013-05-24 07:37:41 -07:00
  • 1a40f936df Windows build cleanup: moving generated lex.cc and parse.cc to build dir Dmitry Babokin 2013-05-20 16:28:20 +04:00
  • 1024ba9b0f Parameter cast for posix_memalign(), otherwise gcc issues an error Dmitry Babokin 2013-05-20 15:30:01 +04:00
  • 1a7ac8b804 Enable memory alignment management via compiler options Dmitry Babokin 2013-05-17 22:15:57 +04:00
  • 7bedb4a081 Add memory alignment dependant on the platform (16/32/64/etc) Dmitry Babokin 2013-05-17 21:35:27 +04:00
  • 630215f56f Defining memory routines completely separately for Windows/Unix 32/64 bit. Dmitry Babokin 2013-05-17 20:13:01 +04:00
  • 6f0e5fd402 Adding RUNTIME define to gen-bitcode-* files generation command Dmitry Babokin 2013-05-17 17:52:21 +04:00
  • 66ec43739a Moving temporary files to Debug/Release folder on Windows Dmitry Babokin 2013-05-17 17:46:32 +04:00
  • 44f9d1ed78 Fix for CYGWIN build warnings (Windows stlye slash, instead of Unix style) Dmitry Babokin 2013-05-17 17:24:42 +04:00
  • c6d479b8ad Enabling 32/64 bit version library build on Windows Dmitry Babokin 2013-05-17 17:08:25 +04:00
  • 80e2f4e342 Removing redundant Debug/Release records in VS proj file - they are unified Dmitry Babokin 2013-05-17 13:03:39 +04:00
  • 4b388edca9 Splitting .ll files to be compiled in two versions - 32 and 64 bit. Unix only Dmitry Babokin 2013-05-17 12:12:03 +04:00
  • 5362dade37 Fixing util.m4 to declare nothing unless some macro is instantiated Dmitry Babokin 2013-05-02 23:22:15 +02:00