aaron/ispc
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Commit Graph

  • eb8e1a2160 +added wc-timer Evghenii 2013-11-08 15:27:51 +01:00
  • ce5f8cd46f replaced with fresh examples Evghenii 2013-11-08 14:17:26 +01:00
  • b2f62d51b0 tuned stencil Evghenii 2013-11-08 14:15:27 +01:00
  • b3c68af40a added volume rendering to run on GPU Evghenii 2013-11-08 13:57:16 +01:00
  • 348100ba42 remove stencil.cubin Evghenii 2013-11-08 10:01:15 +01:00
  • 426afc7377 added workable .cu files for stencil & mandelbrot Evghenii 2013-11-08 10:00:49 +01:00
  • b04caabf39 Merge pull request #656 from egaburov/knc-fix jbrodman 2013-11-07 17:36:36 -08:00
  • 32cfdd52d3 Merge branch 'master' into knc-fix evghenii 2013-11-05 15:46:54 +01:00
  • 015af03bdc changed back to #define ISPC_FORCE_ALIGNED_MEMORY aligned_ld/st #else unaligned ld/st #endif. However load<64>/store<64> will still be unaliged w/o this define because of fails related to the issue #632 evghenii 2013-11-05 15:41:14 +01:00
  • 99946ae0e6 Merge pull request #655 from ifilippov/windows Dmitry Babokin 2013-11-05 05:51:37 -08:00
  • a910bfb539 Windows support Ilia Filippov 2013-10-31 13:25:15 +04:00
  • cb7cbec0d5 added cuda examples Evghenii 2013-11-04 11:44:49 +01:00
  • cb6614da42 fixed the code that non-task code is also emitted for the host Evghenii 2013-11-04 11:33:37 +01:00
  • 6fae459847 a+1 Evghenii 2013-11-04 10:22:05 +01:00
  • bc7c54e5c8 Merge pull request #652 from jbrodman/ptrarithlvalues Dmitry Babokin 2013-11-01 10:39:27 -07:00
  • ab835d8086 Merge pull request #653 from jbrodman/master Dmitry Babokin 2013-11-01 10:39:13 -07:00
  • 0f7050d3aa More stds compliant. VS doesn't like non constant length local arrays. james.brodman 2013-10-31 19:51:13 -04:00
  • 2cef101022 now emits host object file with ptx in it... next step .. testing Evghenii 2013-10-31 18:05:04 +01:00
  • 0a069f7de2 added comment Evghenii 2013-10-31 16:06:44 +01:00
  • dcf9c280ee some cleaning Evghenii 2013-10-31 16:05:06 +01:00
  • a2fd124997 forced module name & ptx string to be generaetd nly once Evghenii 2013-10-31 16:04:30 +01:00
  • 63917f8cc2 now generates CUDALaunch call. Few tweaks are still necesary Evghenii 2013-10-31 16:01:34 +01:00
  • e7ddb9e642 now adds function&module name. next step adding pointer to parameter list Evghenii 2013-10-30 22:41:01 +01:00
  • ec17082864 Add unittest. james.brodman 2013-10-30 17:21:10 -04:00
  • 9ce6fbe1fa Support using pointer arithmetic as lvalue james.brodman 2013-10-30 17:07:26 -04:00
  • 8db3d25844 moved PtxString to Globals Evghenii 2013-10-30 21:05:22 +01:00
  • f9ec1a0097 .. work in programm to embed PTX into host code .. Evghenii 2013-10-30 16:47:30 +01:00
  • 61c33969a2 Merge pull request #651 from jbrodman/shiftfix jbrodman 2013-10-29 11:08:52 -07:00
  • 85eb4cf0d6 Fix logic that looks for shift builtins. james.brodman 2013-10-29 14:02:32 -04:00
  • 47cc470bf6 change nativeVectorWidth from 1 -> 32 for nvptx64 Evghenii 2013-10-29 16:07:12 +01:00
  • 60881499dc Merge branch 'nvptx' of github.com:egaburov/ispc into nvptx egaburov 2013-10-29 15:25:14 +01:00
  • f19cf9274e Merge remote-tracking branch 'upstream/master' into nvptx egaburov 2013-10-29 15:24:40 +01:00
  • ed9bca0e12 add __soa_to_aos*_float1 and __aos_to_soa*_float1 builtins Evghenii 2013-10-29 15:06:08 +01:00
  • f15cdc03e3 nvptx64 generates 2 targets: task and normal function for nvptx64 and export for avx only Evghenii 2013-10-29 14:46:51 +01:00
  • b31fc6f66d now can generate both targets for npvtx64. m_isPTX is set true, to distuish when to either skip or exlcusive euse export Evghenii 2013-10-29 14:17:11 +01:00
  • 8baef6daa3 +1 Evghenii 2013-10-29 14:01:53 +01:00
  • ac700d4860 checkpoint Evghenii 2013-10-29 13:36:31 +01:00
  • b2baa35c3d added correct datalayout for nvptx64 Evghenii 2013-10-29 11:34:01 +01:00
  • b50d3944ea allow easy switch between llvm Evghenii 2013-10-29 10:22:07 +01:00
  • f115a32073 fix llvm 3.2 compilation Evghenii 2013-10-29 10:21:56 +01:00
  • 57aefdf830 accepts ptx extension when target is nvptx64 Evghenii 2013-10-29 10:21:48 +01:00
  • 0d48ace15e Merge pull request #649 from dbabokin/dispatch2 Dmitry Babokin 2013-10-28 14:35:55 -07:00
  • 362ee06b9f Typo fix Dmitry Babokin 2013-10-29 01:35:26 +04:00
  • 9004f090c5 Merge pull request #646 from dbabokin/docs jbrodman 2013-10-28 14:26:17 -07:00
  • 6948120094 Merge pull request #647 from dbabokin/dispatch2 jbrodman 2013-10-28 14:25:41 -07:00
  • 6585a925be Merge pull request #641 from jbrodman/stdlibshift Dmitry Babokin 2013-10-28 14:18:31 -07:00
  • e682b19eda Remove zero initialization for __vec4_i32 james.brodman 2013-10-28 17:13:07 -04:00
  • 8ee3178166 Add Performance Warning james.brodman 2013-10-28 16:51:02 -04:00
  • 09a6e37154 Source cleanup. james.brodman 2013-10-28 16:37:33 -04:00
  • 1b8e745ffe remove condition. Don't use gcc 4.7 for tests. james.brodman 2013-10-28 16:36:59 -04:00
  • 6a1952d1f9 Merge branch 'stdlibshift' of https://github.com/jbrodman/ispc into stdlibshift james.brodman 2013-10-28 16:16:30 -04:00
  • 9ba7b96825 Make the new optimization play nicely with the other.s james.brodman 2013-10-28 16:14:31 -04:00
  • ef5d2dc043 Merge pull request #648 from dbabokin/fail_db Dmitry Babokin 2013-10-28 12:08:30 -07:00
  • 1f0f852fda Standalone checker for detecting the best ISA supported on the host Dmitry Babokin 2013-10-28 22:54:14 +04:00
  • a166eb7ea1 Check AVX OS support in host cpu check code Dmitry Babokin 2013-10-28 22:19:09 +04:00
  • 02681d531e Minor tweak for interface. james.brodman 2013-10-28 12:56:43 -04:00
  • 641d882ea6 Add shift support for knc targets. This is not optimized. james.brodman 2013-10-28 12:43:42 -04:00
  • 1e80b3b0d7 Add shift support for generic-16 target. james.brodman 2013-10-28 12:20:32 -04:00
  • ff98271a43 using mask i1 for nvptx64 Evghenii 2013-10-28 17:03:00 +01:00
  • 500ad7fb51 using mask i1 for nvptx64 Evghenii 2013-10-28 17:01:03 +01:00
  • 4f486333ed now nvptx allows extern "C" task void, which is emits a kernel that should (?) be callable by driver API from external code Evghenii 2013-10-28 16:47:40 +01:00
  • 9a677b62ab Merge branch 'launch3d' into nvptx Evghenii 2013-10-28 14:28:21 +01:00
  • 68ced6ce46 automatically adds -D__NVPTX__ when nvptx64 target is chosen Evghenii 2013-10-28 14:08:32 +01:00
  • 1bd5360d3b added now NVPTX64 automatically emits unmasked extern "C" for task function with kernel attributes Evghenii 2013-10-28 13:58:01 +01:00
  • a7aa1ac1cf now nvptx allows extern "C" task void, which is emits a kernel that should (?) be callable by driver API from external code Evghenii 2013-10-28 12:57:09 +01:00
  • ae23320417 added metadata for tasks with nvptx64 target. not tasks are kernel callable from host Evghenii 2013-10-28 12:10:40 +01:00
  • b68a751f4e generating proper tasking function for nvptx Evghenii 2013-10-28 11:36:08 +01:00
  • 8391d05697 added blockIndex computations Evghenii 2013-10-28 10:18:30 +01:00
  • 63a3214cc6 Removing fails with g++4.4/g++4.7, as we are using clang by default now Dmitry Babokin 2013-10-28 12:45:39 +04:00
  • 4382902894 Fail_db update on Windows: 3.3 update and adding 3.4 Dmitry Babokin 2013-10-28 12:31:24 +04:00
  • 103ef25f12 Docs fix in memory management section Dmitry Babokin 2013-10-27 23:01:20 +04:00
  • 84a7a5d1cb added tests for 3d launch Evghenii 2013-10-26 16:16:28 +02:00
  • ac095dbf3e working on nvptx Evghenii 2013-10-26 16:12:33 +02:00
  • a508bd4290 MacOS fails update Dmitry Babokin 2013-10-26 14:50:45 +04:00
  • c1de753db6 Merge pull request #645 from ifilippov/636 Dmitry Babokin 2013-10-25 12:01:20 -07:00
  • 621679245a fixing problem 644 Ilia Filippov 2013-10-25 12:49:06 +04:00
  • 58aea1b61c Fail_db update with Linux passed with LLVM 3.4 Dmitry Babokin 2013-10-25 21:42:57 +04:00
  • 9b5ee1b31b fail_db update on Linux Dmitry Babokin 2013-10-23 18:42:49 +04:00
  • 383e804ec1 changed notation form taskIndex1,2,3 -> taskIndex0,1,2 Evghenii 2013-10-24 17:20:56 +02:00
  • 34f7986e19 Merge pull request #643 from ifilippov/testing Dmitry Babokin 2013-10-24 05:31:33 -07:00
  • c5fc47cc19 tasksys cleaning egaburov 2013-10-24 14:09:46 +02:00
  • 814ee67519 patch and regression test for problem with vzeroupper Ilia Filippov 2013-10-24 11:51:33 +04:00
  • 43761173ec changed notation, task[Index,Count]_[x,y,z] -> task[Index,Count][1,2,3]. Change launch <<< nx,ny,nz >>> into launch [nx,ny,nz] or equivalent launch [nz][ny][nx]. Programmer can pick the one the is liked the most Evghenii 2013-10-24 13:16:23 +02:00
  • d2b89e0e37 Tweak generic target. james.brodman 2013-10-23 18:01:01 -04:00
  • c4ad8f6ed4 Add docs/generic impls james.brodman 2013-10-23 15:51:59 -04:00
  • 4d289b16c2 Redesign after being hit with the KISS bat. james.brodman 2013-10-23 14:25:43 -04:00
  • e6c8765891 fixed tasksys.cpp for 3d tasking egaburov 2013-10-23 13:18:22 +02:00
  • f89bad1e94 launch now passes the right info into tasking egaburov 2013-10-23 12:51:06 +02:00
  • ade8751442 taskIndex_x,y,z are passed to the task egaburov 2013-10-23 08:39:17 +02:00
  • f97a2d68c8 Bugfix for non-const shift amt and unit tests. james.brodman 2013-10-22 18:29:20 -04:00
  • 899f85ce9c Initial Support for new stdlib shift operator james.brodman 2013-10-22 18:06:54 -04:00
  • 78a05777bc added taskIndex_x,y,z and taskCount_x,y,z egaburov 2013-10-22 16:18:40 +02:00
  • e436e33771 Merge pull request #639 from ifilippov/master Dmitry Babokin 2013-10-21 02:32:05 -07:00
  • faa69dc71f Merge pull request #638 from ifilippov/testing Dmitry Babokin 2013-10-21 02:30:05 -07:00
  • d72590ede6 correction errors in generic targets after operators support Ilia Filippov 2013-10-21 12:35:53 +04:00
  • c378429ffb time in alloy Ilia Filippov 2013-10-18 19:45:39 +04:00
  • ebe1831e71 Merge pull request #636 from ifilippov/master Dmitry Babokin 2013-10-18 03:29:47 -07:00
  • 8cdc70a5b9 Merge pull request #637 from dbabokin/egaburov-avx2-i64x4 Dmitry Babokin 2013-10-18 03:25:10 -07:00
  • 6244902931 Updating fail_db with new Windows fails Dmitry Babokin 2013-10-18 01:16:25 +04:00
  • 1bd5b704c6 Adding support for build on Windows for avx1.1-i64x4 and avx2-i64x4 Dmitry Babokin 2013-10-18 01:15:35 +04:00