{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":550179578,"defaultBranch":"main","name":"pytorch-gpu-benchmark","ownerLogin":"lamikr","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2022-10-12T10:29:31.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/1907638?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1717789618.0","currentOid":""},"activityList":{"items":[{"before":"4bab725956dbe3435f0464c0c5ac8f0c4a1c1395","after":"1401301789dfaff522ea8a7900e96c562892ec91","ref":"refs/heads/main","pushedAt":"2024-08-11T09:39:37.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"update the readme\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"update the readme"}},{"before":"9eec081a52dd836fea86825d4478ceb5a69c6a11","after":"4bab725956dbe3435f0464c0c5ac8f0c4a1c1395","ref":"refs/heads/main","pushedAt":"2024-08-11T09:35:30.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"show benchmark results with new pytorch optimization for 7700S\n\n- latest rocm sdk builder has pytorch optimization in aotriton\n  provided flash attenation algorithm for gfx1100/gfx1101/gfx1102\n  and gfx1103.\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"show benchmark results with new pytorch optimization for 7700S"}},{"before":"e85f6b6acc73b33641e1e46409b41f68383fa58d","after":"9eec081a52dd836fea86825d4478ceb5a69c6a11","ref":"refs/heads/main","pushedAt":"2024-08-03T09:28:03.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"add AMD 680M iGPU gfx1035 results with 6.10.2 kernel\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"add AMD 680M iGPU gfx1035 results with 6.10.2 kernel"}},{"before":"75b52f6e526f349003a598d507c6a13692e52e0b","after":"e85f6b6acc73b33641e1e46409b41f68383fa58d","ref":"refs/heads/main","pushedAt":"2024-08-03T03:25:28.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"readme updates\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"readme updates"}},{"before":"fdaf038e91d072efc0b26f88293ed7642fcf6306","after":"75b52f6e526f349003a598d507c6a13692e52e0b","ref":"refs/heads/main","pushedAt":"2024-08-03T03:13:24.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"small readme update\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"small readme update"}},{"before":"2dee76875f71c5b433136099ff6d46c8ef271669","after":"fdaf038e91d072efc0b26f88293ed7642fcf6306","ref":"refs/heads/main","pushedAt":"2024-08-03T03:02:39.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"add amd framework 16 gpu benchmarks\n\n- add first benchmark results with the\n  framework 16 laptop which has both the\n  discrete 7700S gpu and internal M780M iGPU.\n- tests were executed with rocm sdk builder v6.1.2 version\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"add amd framework 16 gpu benchmarks"}},{"before":"67aa35ef0cf9d227d5e344b79e1374e80eeec3a2","after":"2dee76875f71c5b433136099ff6d46c8ef271669","ref":"refs/heads/main","pushedAt":"2024-08-02T05:29:57.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"initial benchmark result plotter\n\n- datafiles to be plotted must still be\n  hardcoded in the code as a list\n- can show graphs that can be used to compare\n  results from multiple amd and nvidia gpus\n- problem when showing the results is that\n  the benchmarks that have been run on different\n  GPU's have varied during the time. The list of\n  benchmarks in full set is nowadays much bigger\n  than on rx 1080/2060 release time.\n\nSigned-off-by: Mika Laitio <lamikr@pilppa.org>","shortMessageHtmlLink":"initial benchmark result plotter"}},{"before":"eb53c390eb41c3ce49f04991f8219ca292c7dc7a","after":"67aa35ef0cf9d227d5e344b79e1374e80eeec3a2","ref":"refs/heads/main","pushedAt":"2024-07-25T05:28:46.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"use gpu memory to select benchmark models\n\n- defined minimal, medium and full set of models\n- use the amount of gpu memory as a criteria\n  whether to run minimal. medium or full set of models\n  (8GB device can run medium set without getting\n   out of memory error)\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"use gpu memory to select benchmark models"}},{"before":"4d1813fad0689753fc5cb8f02f819902e00a2bb1","after":"eb53c390eb41c3ce49f04991f8219ca292c7dc7a","ref":"refs/heads/main","pushedAt":"2024-07-22T22:47:18.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"updated readme.sh\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"updated readme.sh"}},{"before":"f9f8b05ab3ef35a4f87d568c5933c6a685261b14","after":"4d1813fad0689753fc5cb8f02f819902e00a2bb1","ref":"refs/heads/main","pushedAt":"2024-07-22T06:29:25.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"benchmarks on devices having both amd and nvidia gpus\n\nuse pytorch script to get device count instead\nof using amd or nvidia gpu specific commands and\nthen execute the benchmark for each gpu found.\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"benchmarks on devices having both amd and nvidia gpus"}},{"before":"721915a86f76e06f4a3546db6f236b36ff8f323e","after":"f9f8b05ab3ef35a4f87d568c5933c6a685261b14","ref":"refs/heads/main","pushedAt":"2024-07-21T20:24:00.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"various benchmark execution fixes\n\n- allow specifying gpu-index parameter in addition\n  of gpu-count parameter.\n- gpu index parameter can be used to request the benchmarks\n  to be run only of certain gpu index in multi-gpu case\n- if more than one gpu, run benchmarks separately for each\n  and then in the end run tests with all gpus used at a same time\n- fixes for: https://github.com/lamikr/rocm_sdk_builder/issues/63\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"various benchmark execution fixes"}},{"before":"bff98d04bc73a59f41863045706a09b8de4f5ab6","after":"721915a86f76e06f4a3546db6f236b36ff8f323e","ref":"refs/heads/main","pushedAt":"2024-06-26T23:26:28.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"Add 7900 XTX results\n\n(cherry picked from commit 1d0d0aac9506f9a45ef8ec44bdc18921f2c2176f)","shortMessageHtmlLink":"Add 7900 XTX results"}},{"before":"c221806844dd72f303282f1675b7aff556c4f502","after":"bff98d04bc73a59f41863045706a09b8de4f5ab6","ref":"refs/heads/main","pushedAt":"2024-06-07T19:52:47.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"AMD Radeon RX 6800 gpu benchmark results\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"AMD Radeon RX 6800 gpu benchmark results"}},{"before":"c221806844dd72f303282f1675b7aff556c4f502","after":"bff98d04bc73a59f41863045706a09b8de4f5ab6","ref":"refs/heads/wip/pr_amd_support","pushedAt":"2024-06-07T19:52:40.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"AMD Radeon RX 6800 gpu benchmark results\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"AMD Radeon RX 6800 gpu benchmark results"}},{"before":null,"after":"c221806844dd72f303282f1675b7aff556c4f502","ref":"refs/heads/wip/pr_amd_support","pushedAt":"2024-06-07T19:46:58.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"Add support for running AMD GPU benchmarks\n\n- AMD GPU benchmarks works on rocm platform\n  (tested on rocm 6.1.1) and offers support\n  of using CUDA api with AMD gpus\n- only code change is needed is to test.sh\n  to check whether rocm-smi tool is available\n  and whether it returns AMD GPUs\n- rocm-smi can be used to query the AMD gpu count\n- I had pytorch 2.3.0 and pytorch vision 0.18.0 installed\n  with AMD rocm support enabled and all tests executed\n  without errors on AMD RX 6800\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"Add support for running AMD GPU benchmarks"}},{"before":"a60299c656f837b1a6d633b193cbeec01028170f","after":"c221806844dd72f303282f1675b7aff556c4f502","ref":"refs/heads/main","pushedAt":"2024-06-07T19:43:11.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"Add support for running AMD GPU benchmarks\n\n- AMD GPU benchmarks works on rocm platform\n  (tested on rocm 6.1.1) and offers support\n  of using CUDA api with AMD gpus\n- only code change is needed is to test.sh\n  to check whether rocm-smi tool is available\n  and whether it returns AMD GPUs\n- rocm-smi can be used to query the AMD gpu count\n- I had pytorch 2.3.0 and pytorch vision 0.18.0 installed\n  with AMD rocm support enabled and all tests executed\n  without errors on AMD RX 6800\n\nSigned-off-by: Mika Laitio <lamikr@gmail.com>","shortMessageHtmlLink":"Add support for running AMD GPU benchmarks"}},{"before":"ac5bc3dc3ec8dcaea6273d82a2d2a24145c39ed6","after":"a60299c656f837b1a6d633b193cbeec01028170f","ref":"refs/heads/main","pushedAt":"2024-06-07T19:36:53.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"Update code formatting","shortMessageHtmlLink":"Update code formatting"}},{"before":null,"after":"ac5bc3dc3ec8dcaea6273d82a2d2a24145c39ed6","ref":"refs/heads/amd_gpu_pytorch_20_support_old","pushedAt":"2024-06-07T19:31:06.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"renamed script used to run gpu benchmarks\n\nSigned-off-by: Mika Laitio <lamikr@pilppa.org>","shortMessageHtmlLink":"renamed script used to run gpu benchmarks"}},{"before":"d0208ec2046b4a82197311ad53740ac5f2a3299d","after":"ac5bc3dc3ec8dcaea6273d82a2d2a24145c39ed6","ref":"refs/heads/main","pushedAt":"2024-05-26T23:08:44.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"lamikr","name":"Mika Laitio","path":"/lamikr","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/1907638?s=80&v=4"},"commit":{"message":"renamed script used to run gpu benchmarks\n\nSigned-off-by: Mika Laitio <lamikr@pilppa.org>","shortMessageHtmlLink":"renamed script used to run gpu benchmarks"}}],"hasNextPage":false,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOC0xMVQwOTozOTozNy4wMDAwMDBazwAAAASXUJR2","startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOC0xMVQwOTozOTozNy4wMDAwMDBazwAAAASXUJR2","endCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wNS0yNlQyMzowODo0NC4wMDAwMDBazwAAAARUXKjQ"}},"title":"Activity · lamikr/pytorch-gpu-benchmark"}