{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":590563289,"defaultBranch":"master","name":"openEMS","ownerLogin":"biergaizi","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2023-01-18T17:44:26.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/1310693?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1698018909.0","currentOid":""},"activityList":{"items":[{"before":null,"after":"678495372ea64bd322c7ea48cb6b6ff737cb64a8","ref":"refs/heads/no-ext-no-opcomp","pushedAt":"2023-10-22T23:55:09.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Disable operator compression.\n\nFor testing pure memory bandwidth scalability, there's a need to\ndisble operator compression in the \"SSE Compressed\" engine.","shortMessageHtmlLink":"Disable operator compression."}},{"before":null,"after":"db3300e1c4f44a28e68ad1b5795060621df9477a","ref":"refs/heads/kokkos-staging","pushedAt":"2023-10-22T23:53:26.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Disable operator compression.\n\nFor testing pure memory bandwidth scalability, there's a need to\ndisble operator compression in the \"SSE Compressed\" engine.","shortMessageHtmlLink":"Disable operator compression."}},{"before":"5e334131934a3a53f16e1a32e5bf6245d2031b97","after":"10fad85d3e1fe5ed837d40920ec2692b7beeefb9","ref":"refs/heads/sycl-test1","pushedAt":"2023-07-19T02:22:38.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"more staging (and broken) SYCL code","shortMessageHtmlLink":"more staging (and broken) SYCL code"}},{"before":"56fdf3c87fce09583d2ac1e2e2d691b85a3d0515","after":"5e334131934a3a53f16e1a32e5bf6245d2031b97","ref":"refs/heads/sycl-test1","pushedAt":"2023-07-17T16:07:23.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"more SYCL staging code\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"more SYCL staging code"}},{"before":null,"after":"56fdf3c87fce09583d2ac1e2e2d691b85a3d0515","ref":"refs/heads/sycl-test1","pushedAt":"2023-07-16T17:56:34.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"WIP sycl code","shortMessageHtmlLink":"WIP sycl code"}},{"before":null,"after":"606e71c241fef8a602999934d40d3b032e5d464c","ref":"refs/heads/project-diamond-test1","pushedAt":"2023-07-12T16:33:24.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert compressed operator's lookup table to Flat_3DArray.\n\nThis commit converts the compressed operator's lookup table\nm_Op_index to use Flat_3DArray instead of an array of pointers.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert compressed operator's lookup table to Flat_3DArray."}},{"before":"25ad78b3430980af008fc52523f0f187533f2829","after":"a8a194af205d12463b3b2068b9b046de27a16767","ref":"refs/heads/project-diamond-staging3","pushedAt":"2023-07-06T08:03:16.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"engine_sse: libdivide for fast division to find vectors.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"engine_sse: libdivide for fast division to find vectors."}},{"before":"f3c08e2e3f670b5e8bcda42d0878abb1e72c47fe","after":"25ad78b3430980af008fc52523f0f187533f2829","ref":"refs/heads/project-diamond-staging3","pushedAt":"2023-07-05T15:40:33.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Try 8-element vectorization (aka AVX).\n\nAs far as I can see, there's absolutely no improvement when running\nsingle-threaded (as the bottleneck is memory latency and bandwidth,\nnot CPU throughput), and there's even a huge performance regression\nwhen multi-threaded, probably due to cacheline conflicts between\ncores. But the patch is submitted here for further experiments.","shortMessageHtmlLink":"Try 8-element vectorization (aka AVX)."}},{"before":null,"after":"f3c08e2e3f670b5e8bcda42d0878abb1e72c47fe","ref":"refs/heads/project-diamond-staging3","pushedAt":"2023-07-04T05:49:04.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"tools/tiling.cpp: sync with test repo","shortMessageHtmlLink":"tools/tiling.cpp: sync with test repo"}},{"before":"6fd29deb18c3db218bcfaf9493f2b0d1035104b7","after":"289fef96103ff839c5ad6e56f0119351d0fb055b","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-04T04:27:52.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"engine_tiling.cpp: print the number of tiles in a stage.","shortMessageHtmlLink":"engine_tiling.cpp: print the number of tiles in a stage."}},{"before":"2ed491fecd07094b0406eb396f403238a34f08e2","after":"6fd29deb18c3db218bcfaf9493f2b0d1035104b7","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-01T13:36:25.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"tools/TileMap.h: fix missing std::array include\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"tools/TileMap.h: fix missing std::array include"}},{"before":"66456746b744f9536c5de36f9a1cf8ef36cb5a03","after":"2ed491fecd07094b0406eb396f403238a34f08e2","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-01T05:52:58.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"openems.cpp: use Tiling engine by default (hack).\n\nThis commit is a temporary hack to use the spatial/temporal\nTiling engine by default. Since not all features and extentions\nare supported, it's not a suitable default. However, the Python\nAPI lacks the ability to select the underlying engine, so the\nnew Tiling engine is temporary used by default for testing.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"openems.cpp: use Tiling engine by default (hack)."}},{"before":"492bf483f0bce459ae205a5623e917e0a8741178","after":"66456746b744f9536c5de36f9a1cf8ef36cb5a03","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-01T04:39:02.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"openems.cpp: use Tiling engine by default (hack).\n\nThis commit is a temporary hack to use the spatial/temporal\nTiling engine by default. Since not all features and extentions\nare supported, it's not a suitable default. However, the Python\nAPI lacks the ability to select the underlying engine, so the\nnew Tiling engine is temporary used by default for testing.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"openems.cpp: use Tiling engine by default (hack)."}},{"before":"bcd396a06d2c809e0bf3841a25a9b116371e0c41","after":"492bf483f0bce459ae205a5623e917e0a8741178","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-01T04:25:36.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"openems.cpp: use Tiling engine by default (hack).\n\nThis commit is a temporary hack to use the spatial/temporal\nTiling engine by default. Since not all features and extentions\nare supported, it's not a suitable default. However, the Python\nAPI lacks the ability to select the underlying engine, so the\nnew Tiling engine is temporary used by default for testing.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"openems.cpp: use Tiling engine by default (hack)."}},{"before":"8e331c149f263dcfa54533942924dad475e5fd58","after":"bcd396a06d2c809e0bf3841a25a9b116371e0c41","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-07-01T03:35:54.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Introduce EngineType_Tiling.\n\nThis commit introduces the tiling engine. This is a high-performance\nmulti-thread engine with spatial and temporal tiling. Initial test\nsuggests it achieves a 2x to 6x speedup in many simulation setups.\nThe disadvantage is that not all extensions are supported (in fact,\nit's possible that the cylindrical coordinate system may never be\nsupported), and the correctness has not been full verified.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"Introduce EngineType_Tiling."}},{"before":null,"after":"8e331c149f263dcfa54533942924dad475e5fd58","ref":"refs/heads/project-diamond-rework1","pushedAt":"2023-06-30T08:32:49.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Excitation: add support for tiling.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"Excitation: add support for tiling."}},{"before":"794efdfdfe5869ed621d7ab41ec004ce4358b9c2","after":"8eca218e97c124014e6f1054e77d7baa92af816b","ref":"refs/heads/project-diamond-staging2","pushedAt":"2023-06-30T01:58:29.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"more staging changes","shortMessageHtmlLink":"more staging changes"}},{"before":null,"after":"794efdfdfe5869ed621d7ab41ec004ce4358b9c2","ref":"refs/heads/project-diamond-staging2","pushedAt":"2023-06-26T07:36:31.045Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"staging for true diamond tiling","shortMessageHtmlLink":"staging for true diamond tiling"}},{"before":null,"after":"b9f82500f00a985b1bc7091c3a52eb526d0ef7a4","ref":"refs/heads/project-diamond","pushedAt":"2023-06-25T04:57:18.831Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"Proper cleanup","shortMessageHtmlLink":"Proper cleanup"}},{"before":"77d042d15a1b4672a854871b259d42992eb60305","after":null,"ref":"refs/heads/rework-stage3-xyzn-pad","pushedAt":"2023-05-20T14:11:18.665Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"}},{"before":null,"after":"77d042d15a1b4672a854871b259d42992eb60305","ref":"refs/heads/rework-stage3-corrected-nxyz-pad","pushedAt":"2023-05-20T14:11:18.288Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":"da7a4541aace747ac8800e073be2ce048a981e1d","after":null,"ref":"refs/heads/rework-stage3-nxyz","pushedAt":"2023-05-20T14:11:02.835Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"}},{"before":null,"after":"da7a4541aace747ac8800e073be2ce048a981e1d","ref":"refs/heads/rework-stage3-corrected-xyzn","pushedAt":"2023-05-20T14:11:02.460Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":"373766d9e30af1dae2e6677ac1967385882d497a","after":"77d042d15a1b4672a854871b259d42992eb60305","ref":"refs/heads/rework-stage3-xyzn-pad","pushedAt":"2023-05-20T14:09:58.294Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":"40029fdf021f9c0a829ebe03d4ddbebdcdded66e","after":"da7a4541aace747ac8800e073be2ce048a981e1d","ref":"refs/heads/rework-stage3-nxyz","pushedAt":"2023-05-20T14:09:36.691Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":"c4686c4c9cb990fc65b1c8bc85632eb6ba055b44","after":null,"ref":"refs/heads/rework-stage3-corrected-xyzn","pushedAt":"2023-05-20T14:09:03.848Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"}},{"before":"1b53d981e3ea260cb6bc817e8881715e97510ac0","after":null,"ref":"refs/heads/rework-stage3-corrected-nxyz-pad","pushedAt":"2023-05-20T14:08:52.488Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"}},{"before":null,"after":"1b53d981e3ea260cb6bc817e8881715e97510ac0","ref":"refs/heads/rework-stage3-corrected-nxyz-pad","pushedAt":"2023-05-20T13:39:58.522Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":null,"after":"c4686c4c9cb990fc65b1c8bc85632eb6ba055b44","ref":"refs/heads/rework-stage3-corrected-xyzn","pushedAt":"2023-05-20T13:38:41.668Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"FDTD: convert UPML to use Flat_N_3DArray.\n\nThe performance improvement is around 15% in addition to the\nprevious 15% speedup of the FDTD kernel, allowing an up-to-30%\nspeedup in many simulations.\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"FDTD: convert UPML to use Flat_N_3DArray."}},{"before":"ed191ad520098f6bfd25143178a8bb842c6aa4df","after":"7d1960d8eb1edd48df214578033379e9a43a31e1","ref":"refs/heads/preserve-cflags","pushedAt":"2023-05-16T04:55:29.467Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"biergaizi","name":"biergaizi","path":"/biergaizi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1310693?s=80&v=4"},"commit":{"message":"CMakeLists.txt: append instead of overwrite CMAKE_CXX_FLAGS\n\nCurrently, on ARM and PPC, CMakeLists.txt uses:\n\n    set(CMAKE_CXX_FLAGS \"-DNO_WARN_X86_INTRINSICS -DSSE_CORRECT_DENORMALS\")\n\nbut this overwrites the default value of CMAKE_CXX_FLAGS from CMake,\nincluding user-specified CXXFLAGS via environmental variable, making\nit impossible to change CXXFLAGS.\n\nThis patch appends instead of overwrite CMAKE_CXX_FLAGS via:\n\n    set(CMAKE_CXX_FLAGS \"${CMAKE_CXX_FLAGS} -DNO_WARN_X86_INTRINSICS -DSSE_CORRECT_DENORMALS\")\n\nSigned-off-by: Yifeng Li <tomli@tomli.me>","shortMessageHtmlLink":"CMakeLists.txt: append instead of overwrite CMAKE_CXX_FLAGS"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAADnR9ecAA","startCursor":null,"endCursor":null}},"title":"Activity · biergaizi/openEMS"}