{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":193269193,"defaultBranch":"master","name":"pytorch","ownerLogin":"andrewjcg","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2019-06-22T18:54:19.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/648225?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1718117143.0","currentOid":""},"activityList":{"items":[{"before":null,"after":"8a75c87183005f9b610e3e96260d4851963bff09","ref":"refs/heads/python_rel_lib_path","pushedAt":"2024-06-11T14:45:43.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"Allow setting `PYTHON_LIB_REL_PATH` via environment variable\n\nThis allows builds to customize the location where caffe2's Python\nmodules are installed to.","shortMessageHtmlLink":"Allow setting PYTHON_LIB_REL_PATH via environment variable"}},{"before":null,"after":"0452abd4bd83f3e871dcd0b4866b5b93a7574d67","ref":"refs/heads/viable/strict","pushedAt":"2024-06-07T02:18:39.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"fix","shortMessageHtmlLink":"fix"}},{"before":"fc3bee93f4ec17fb293971857b1710e16bbb51a3","after":"3d5542bd5b47ca1c49ed91a0d2360edaaa61ea2a","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-04T15:23:50.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"e5a2a4499376d5cf6f181981285412e6b24ecfae","after":"fc3bee93f4ec17fb293971857b1710e16bbb51a3","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-04T14:40:23.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"2ff2271acc39d9c0fa454c855fa5ad6c8eee992f","after":"e5a2a4499376d5cf6f181981285412e6b24ecfae","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-04T14:31:05.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"fbf74854c55752fc00a12c239f74db8ee437af2b","after":"2ff2271acc39d9c0fa454c855fa5ad6c8eee992f","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-03T21:36:43.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"383d03e0a474aefaaaf17d36391391627a0e4eb4","after":"fbf74854c55752fc00a12c239f74db8ee437af2b","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-03T21:22:04.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"2ecdb45f085560c9473975606d680d626ed53ac4","after":"383d03e0a474aefaaaf17d36391391627a0e4eb4","ref":"refs/heads/export-D55206000","pushedAt":"2024-04-03T21:13:44.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":"3e53af9f09640eacd872ef7ed9bfe2b4bb14cffd","after":"2ecdb45f085560c9473975606d680d626ed53ac4","ref":"refs/heads/export-D55206000","pushedAt":"2024-03-22T15:16:31.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433 (#122470)\n\nSummary:\n\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433 (pytorch#122470)"}},{"before":null,"after":"3e53af9f09640eacd872ef7ed9bfe2b4bb14cffd","ref":"refs/heads/export-D55206000","pushedAt":"2024-03-22T02:36:51.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[triton] Backport https://github.com/openai/triton/pull/3433\n\nSummary:\nPull cache API changes from https://github.com/openai/triton/pull/3433.\n\nAmong other simplifications, this allows us the cache all files in a \"group\"\natomically, in a single memcache blob, and avoid needing to use other\napproaches to handle these files coming from different runs.\n\nContext:\nhttps://fb.workplace.com/groups/420659799592399/posts/778155640509478/\n\nScuba query for these changes:\nhttps://fburl.com/scuba/triton_remote_cache/lb3t1cw4\n\nTest Plan:\nWith D55206078:\n\n```\n$ TORCHBENCH_TOL='1e-3' TORCHINDUCTOR_PERMUTE_FUSION='1' TORCHINDUCTOR_SHAPE_PADDING='1' buck2 run mode/opt //pytorch/benchmark:run -- cmf_10x -d cuda -t train --torchdynamo inductor\n```\n\nReviewed By: bertmaher\n\nDifferential Revision: D55206000","shortMessageHtmlLink":"[triton] Backport triton-lang/triton#3433"}},{"before":null,"after":"7e5c005f02eaa5e42c4cde3136213f3df60f6292","ref":"refs/heads/export-D51444766","pushedAt":"2023-11-18T18:12:03.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2] Add non-x86 stub definition for `libraryFor` too\n\nSummary: Fix non-x86 build errors with missing `libraryFor` symbol.\n\nTest Plan:\n```\n$ buck2 build -c fbcode.arch=aarch64 fbcode//admarket/adfinder:adfinder\n```\n\nReviewed By: malfet\n\nDifferential Revision: D51444766","shortMessageHtmlLink":"[caffe2] Add non-x86 stub definition for libraryFor too"}},{"before":"14dba0447d45f43856ca8011d5362b86fa2180d7","after":"9baa08c5cacf73d8058c17d2f902eb4b6f8227a4","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T03:36:26.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"f2899c342dcecd82544fcc6b5d5a50954176679b","after":"14dba0447d45f43856ca8011d5362b86fa2180d7","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T03:35:59.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"be38f0da505b02c92bb6cda3a138a6b2b3ea7874","after":"f2899c342dcecd82544fcc6b5d5a50954176679b","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T01:59:05.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"e4c19d6b779cb744a84ba01b7688227fa1b00e8a","after":"be38f0da505b02c92bb6cda3a138a6b2b3ea7874","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T01:59:04.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"bc884366ecbc506e5badb2796d10079eba03bf70","after":"e2f1a8a0a325926950e8900b5f2dca983ed6cfd2","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T01:58:04.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"e2f1a8a0a325926950e8900b5f2dca983ed6cfd2","after":"e4c19d6b779cb744a84ba01b7688227fa1b00e8a","ref":"refs/heads/export-D47242468","pushedAt":"2023-09-15T01:58:04.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly (#104707)\n\nSummary:\n\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park, aaronenyeshi\n\nDifferential Revision: D47242468","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assem…"}},{"before":"ee4b578c1d85316aab8ae8235d534fd88e453636","after":"4d444846f205b3a22b2f894495beda9a19976cd6","ref":"refs/heads/export-D47844555","pushedAt":"2023-07-28T17:56:01.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2] Clean up platform-specific fbobjc deps/flags (#106193)\n\nSummary:\nPull Request resolved: https://github.com/pytorch/pytorch/pull/106193\n\nReplace `platform_deps` with `select()`s and use to make sure the `cpukernel_avx2`\ndep is x86-specific.\n\nhttps://fb.prod.workplace.com/groups/buck2users/posts/3469961626593527/\n\nTest Plan:\n```\n$ buck2 build //xplat/rtc/media/tools/newton:newton_pcAppleMac#macosx-arm64 --target-platforms //xplat/rtc/webrtc/platforms:bwe-dbg-arm64\n```\n\nDifferential Revision: D47844555\n\nfbshipit-source-id: c300a37ce924f198b6c0ef7c20c1ea78d9458fe0","shortMessageHtmlLink":"[caffe2] Clean up platform-specific fbobjc deps/flags (pytorch#106193)"}},{"before":null,"after":"ee4b578c1d85316aab8ae8235d534fd88e453636","ref":"refs/heads/export-D47844555","pushedAt":"2023-07-28T05:41:47.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2] Clean up platform-specific fbobjc deps/flags\n\nSummary:\nReplace `platform_deps` with `select()`s and use to make sure the `cpukernel_avx2`\ndep is x86-specific.\n\nhttps://fb.prod.workplace.com/groups/buck2users/posts/3469961626593527/\n\nTest Plan:\n```\n$ buck2 build //xplat/rtc/media/tools/newton:newton_pcAppleMac#macosx-arm64 --target-platforms //xplat/rtc/webrtc/platforms:bwe-dbg-arm64\n```\n\nDifferential Revision: D47844555\n\nfbshipit-source-id: ac4c16140c304a0aa4b3a8cf0b2765a01b6169a4","shortMessageHtmlLink":"[caffe2] Clean up platform-specific fbobjc deps/flags"}},{"before":null,"after":"0e9ef3135a71da94bc9b2b12e7f99b39fdc76842","ref":"refs/heads/export-D47772650","pushedAt":"2023-07-26T01:52:42.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][xplat/caffe2] Fix aarch64 build\n\nSummary: Add missing dep to pull in sleef headers.\n\nTest Plan:\n```\n$ buck2 build -c fbcode.arch=aarch64 fbsource//xplat/caffe2:cpukernel_avx2\n```\n\nReviewed By: r1mikey, ajaymh\n\nDifferential Revision: D47772650\n\nfbshipit-source-id: 6e50d764bc231eea3e60cbad0f20b24c02f95cce","shortMessageHtmlLink":"[aarch64][xplat/caffe2] Fix aarch64 build"}},{"before":null,"after":"bc884366ecbc506e5badb2796d10079eba03bf70","ref":"refs/heads/export-D47242468","pushedAt":"2023-07-06T14:38:08.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly\n\nSummary:\nPort x86 inline assembly to aarch64:\n- Use `sp` instead of `%rsp` for stack pointer; move to second caller-\n saved register `x1` instead of `%rsi`\n- Use `x29` instead of `%rbp` for base pointer; move to third caller-\n saved register `x2` instead of `%rdx`\n\nTest Plan:\n```\n$ buck2 build fbcode//mode/opt fbcode//caffe2/torch/fb/model_transform/fx2trt/packaging:generate_merge_net_file\n```\n\nReviewed By: jasonjk-park\n\nDifferential Revision: D47242468\n\nfbshipit-source-id: bb3cb81d866e27c1f19c4d9da9a8ed57e21362c2","shortMessageHtmlLink":"[aarch64][caffe2/torch/csrc/profiler] Support aarch64 in inline assembly"}},{"before":"f7d47df7406c9319384d089ffec0ddd8d505ca51","after":"f63a6b1b9103b2ec53358dfc0a79e90c8dfa4d9d","ref":"refs/heads/export-D47113122","pushedAt":"2023-07-02T16:42:56.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[PyTorch][Dispatcher] Fix destruction order fiasco crash (#104393)\n\nSummary:\nPull Request resolved: https://github.com/pytorch/pytorch/pull/104393\n\nThe current implementation of `Dispatcher` returns an RAII object\nfrom it's `register*` methods which, on destruction, uses a saved\nreference to the `Dispatcher` to call the associated `deregister*`\nmethod.\n\nHowever, nothing guarantees that the `Dispatcher` is destroyed\n*after* all RAII objects have been destroyed and, in practice, we\nsee segfaults caused when a global `Dispatcher` is cleaned up\nbefore RAII globals.\n\nThis diff fixes by keeping the `Dispatcher` lock and \"alive\" marker\nin a `std::shared_ptr` which the callbacks copy and then use to\nverify that the `Dispatcher` is still alive before continuing.\n\nhttps://fb.workplace.com/groups/1405155842844877/posts/7143161099044294/\nhttps://fb.workplace.com/groups/python.builds/posts/3510588832595867/\nS349108\n\nTest Plan: CI\n\nReviewed By: BrandonTheBuilder\n\nDifferential Revision: D47113122\n\nfbshipit-source-id: 84d3be783dc6a0f3f9e2a68a7375881a2f9b64f3","shortMessageHtmlLink":"[PyTorch][Dispatcher] Fix destruction order fiasco crash (pytorch#104393"}},{"before":null,"after":"f7d47df7406c9319384d089ffec0ddd8d505ca51","ref":"refs/heads/export-D47113122","pushedAt":"2023-06-29T02:35:28.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[PyTorch][Dispatcher] Fix destruction order fiasco crash\n\nSummary:\nThe current implementation of `Dispatcher` returns an RAII object\nfrom it's `register*` methods which, on destruction, uses a saved\nreference to the `Dispatcher` to call the associated `deregister*`\nmethod.\n\nHowever, nothing guarantees that the `Dispatcher` is destroyed\n*after* all RAII objects have been destroyed and, in practice, we\nsee segfaults caused when a global `Dispatcher` is cleaned up\nbefore RAII globals.\n\nThis diff fixes by keeping the `Dispatcher` lock and \"alive\" marker\nin a `std::shared_ptr` which the callbacks copy and then use to\nverify that the `Dispatcher` is still alive before continuing.\n\nhttps://fb.workplace.com/groups/1405155842844877/posts/7143161099044294/\nhttps://fb.workplace.com/groups/python.builds/posts/3510588832595867/\nS349108\n\nTest Plan: CI\n\nDifferential Revision: D47113122\n\nfbshipit-source-id: bc2d23f5a9082818ff0f65e38a392ab2860eb678","shortMessageHtmlLink":"[PyTorch][Dispatcher] Fix destruction order fiasco crash"}},{"before":"4f9888f3ffb01925b66971c04e046ec568464f46","after":"4ae74d5186ff5ae265629ebb9894f6e535e13c62","ref":"refs/heads/nondeterm_autograd2","pushedAt":"2023-05-15T21:22:30.362Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2/tools/autograd] Fix non-determinism in code gen\n\nSummary:\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code.","shortMessageHtmlLink":"[caffe2/tools/autograd] Fix non-determinism in code gen"}},{"before":null,"after":"4f9888f3ffb01925b66971c04e046ec568464f46","ref":"refs/heads/nondeterm_autograd2","pushedAt":"2023-05-15T17:50:08.091Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2/tools/autograd] Fix non-determinism in code gen\n\nSummary:\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code.","shortMessageHtmlLink":"[caffe2/tools/autograd] Fix non-determinism in code gen"}},{"before":"d15cdab434d36c9eec52ee75f740c97b16d22844","after":"626df7adcd0c98dbb0c3671e8799e2e1dd751836","ref":"refs/heads/nondeterm_torchgen","pushedAt":"2023-05-15T15:37:29.100Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"pytorchmergebot","name":null,"path":"/pytorchmergebot","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/97764156?s=80&v=4"},"commit":{"message":"[caffe2/torchgen] Fix codegen non-determinism (#100986)\n\nSummary:\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code.","shortMessageHtmlLink":"[caffe2/torchgen] Fix codegen non-determinism (pytorch#100986)"}},{"before":"0d1a056c4e9b18e0e0dec2e81bbe014202da765e","after":"d15cdab434d36c9eec52ee75f740c97b16d22844","ref":"refs/heads/nondeterm_torchgen","pushedAt":"2023-05-15T15:26:55.544Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2/torchgen] Fix codegen non-determinism (#100986)\n\nSummary:\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code.","shortMessageHtmlLink":"[caffe2/torchgen] Fix codegen non-determinism (pytorch#100986)"}},{"before":"f2a6a18f7dc0f05c5e0a17138b1ca4ce58d50302","after":"0d1a056c4e9b18e0e0dec2e81bbe014202da765e","ref":"refs/heads/nondeterm_torchgen","pushedAt":"2023-05-14T15:48:34.535Z","pushType":"push","commitsCount":1,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"Update torchgen/gen.py\n\nCo-authored-by: Aaron Gokaslan ","shortMessageHtmlLink":"Update torchgen/gen.py"}},{"before":"2ce0771ddf2381177eb51b69c345153136a7e825","after":"4f9888f3ffb01925b66971c04e046ec568464f46","ref":"refs/heads/nondeterm_autograd","pushedAt":"2023-05-12T17:09:17.063Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2/tools/autograd] Fix non-determinism in code gen\n\nSummary:\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code.","shortMessageHtmlLink":"[caffe2/tools/autograd] Fix non-determinism in code gen"}},{"before":null,"after":"2ce0771ddf2381177eb51b69c345153136a7e825","ref":"refs/heads/nondeterm_autograd","pushedAt":"2023-05-12T16:26:11.190Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"andrewjcg","name":null,"path":"/andrewjcg","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/648225?s=80&v=4"},"commit":{"message":"[caffe2/tools/autograd] Fix non-determinism in code gen (#100992)\n\nSummary:\nPull Request resolved: https://github.com/pytorch/pytorch/pull/100992\n\nFix several cases of leaking set-iteration-order to generated\nsources, causing non-determinism in generated code from\n`fbcode//caffe2:generate-code`, which was causing spurious\nre-compilation of `fbcode//caffe2:libtorch`.","shortMessageHtmlLink":"[caffe2/tools/autograd] Fix non-determinism in code gen (pytorch#100992)"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEYljCMgA","startCursor":null,"endCursor":null}},"title":"Activity · andrewjcg/pytorch"}