{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":37176333,"defaultBranch":"main","name":"ompi","ownerLogin":"bosilca","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2015-06-10T04:57:20.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/642701?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1724337780.0","currentOid":""},"activityList":{"items":[{"before":"ce2310addf605c71afd69ec8c1108eb7eb155585","after":"5f00259f9efd5b88bb178ba0070262175daa519b","ref":"refs/heads/main","pushedAt":"2024-09-18T23:36:45.000Z","pushType":"push","commitsCount":82,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12783 from amd-nithyavs/29Aug2024_nocid_array\n\ncoll/acoll: Remove use of cid as array index","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2494896642\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12783\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12783/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12783\">open-mpi#12783</a> from amd-nithyavs/29Aug2024_nocid_a…"}},{"before":null,"after":"6c5c69e07cf9e154c926e58c350183183120fa24","ref":"refs/heads/topic/no_btl_smcuda_without_accelerator","pushedAt":"2024-08-22T14:43:00.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Deal with PMIX optional keys\n\nPMIX returns success and does not set the output to NULL while looking\nfor an optional key. Thus, to prevent segfaults we need to set the\noutput value to a known value before.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Deal with PMIX optional keys"}},{"before":"0a0249117ef22c6a8df6a69a2d4a78236002a1f0","after":null,"ref":"refs/heads/topic/fix_same_instance","pushedAt":"2024-07-25T15:00:42.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"}},{"before":null,"after":"0a0249117ef22c6a8df6a69a2d4a78236002a1f0","ref":"refs/heads/topic/fix_same_instance","pushedAt":"2024-07-25T04:02:22.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Fix the access to the rcount array.\n\nSomehow this has been missed during the MPI_Count upgrade of the\ncollective API.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Fix the access to the rcount array."}},{"before":"ff12b69cdd4b09871accd7ac7cd6870f0a55ecb3","after":"ce2310addf605c71afd69ec8c1108eb7eb155585","ref":"refs/heads/main","pushedAt":"2024-07-25T03:31:41.000Z","pushType":"push","commitsCount":37,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12681 from ggouaillardet/topic/nvfortran_ignore_tkr\n\nconfigury: harden IGNORE_TKR check","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2407823040\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12681\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12681/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12681\">open-mpi#12681</a> from ggouaillardet/topic/nvfortran_…"}},{"before":"04808815237e422c799502fe27a844dda797b248","after":"7a4e296ac53b2f351dc14ab05d8b2a6c9b24fe7f","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-24T04:37:00.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Add accelerator support for packing and unpacking\n\nSupport all types of datatype, including vector, indexed and struct in\naddition to contiguous. The packing is done by a kernel that is\ncurrently launched for each operation.\n\nBasically, the datatype description is transposed into a flatten iovec,\nand this list of iovecs is i[evenly] distributed amoung all the SM used\nfor the kernel.\n\nAdd a tests to validate the packing/unpacking.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Add accelerator support for packing and unpacking"}},{"before":null,"after":"c3498bcf3d263e5985c18a308806ccc1a5d87ade","ref":"refs/heads/topic/port/12691","pushedAt":"2024-07-19T21:21:33.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Check that requests belong to similar sessions\n\nConsistent way to check that an array of requests belong to the same\nsession or are predefined. Skip all requests generated by components\nthat do not support instances (generalized requests, file, win).\n\nReplaces #12525.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>\n(cherry picked from commit ae5b0f6bc7dbb04f0d1c2c999c2275f779c3ba67)","shortMessageHtmlLink":"Check that requests belong to similar sessions"}},{"before":"ae5b0f6bc7dbb04f0d1c2c999c2275f779c3ba67","after":null,"ref":"refs/heads/topic/request_single_instance_fix","pushedAt":"2024-07-18T18:43:27.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"}},{"before":null,"after":"ae5b0f6bc7dbb04f0d1c2c999c2275f779c3ba67","ref":"refs/heads/topic/request_single_instance_fix","pushedAt":"2024-07-18T17:01:27.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Check that requests belong to similar sessions\n\nConsistent way to check that an array of requests belong to the same\nsession or are predefined. Skip all requests generated by components\nthat do not support instances (generalized requests, file, win).\n\nReplaces #12525.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Check that requests belong to similar sessions"}},{"before":"f117f2637752be3de860aae292aedc5ad22df3fa","after":"04808815237e422c799502fe27a844dda797b248","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-06T03:00:04.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Add accelerator support for packing and unpacking\n\nSupport all types of datatype, including verctor, indexed and struct in\naddition to contiguous. The packing is done by a kernel that is\ncurrently launched for each operation.\n\nBasically, the datatype description is transposed into a flatten iovec,\nand this list of iovecs is i[evenly] distributed amoung all the SM used\nfor the kernel.\n\nAdd a tests to validate the packing/unpacking.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Add accelerator support for packing and unpacking"}},{"before":"02fce5c281e7495470b336d0166d503762302c5c","after":"f117f2637752be3de860aae292aedc5ad22df3fa","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-06T02:38:00.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Add accelerator support for packing and unpacking\n\nSupport all types of datatype, including verctor, indexed and struct in\naddition to contiguous. The packing is done by a kernel that is\ncurrently launched for each operation.\n\nBasically, the datatype description is transposed into a flatten iovec,\nand this list of iovecs is i[evenly] distributed amoung all the SM used\nfor the kernel.\n\nAdd a tests to validate the packing/unpacking.\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Add accelerator support for packing and unpacking"}},{"before":"6f31034d2d042309b8081bf243e3d3b00d2e2b31","after":"02fce5c281e7495470b336d0166d503762302c5c","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-05T22:37:51.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Add GPU packing and unpacking\nadd cuda stream for submmitting multiple kernels.\nadd suppot for predefined datatypes.\n\nindexed datatype new, bonus stask support.\nAdd support for iovec and for pipeline iovec.\na new way to compute nb_block and thread_per_block\n\nRDMA send is now working.\n\nAdd support for vector datatype. Add pipeline.\nImprove the GPU memory management.\n\nfix gpu memory and vector datatype\n\nunrestricted GPU. Instead of forcing everything to go on\ndevice 0, we now use the devices already opened.\n\nUsing globally defined indexes lead to several synchronization\nissues, when 2 peers were doing a send/recv or when multiple senders\nwere targetting the same receiver. Rolf provided a patch to solve\nthis issue, by moving the IPC communication index from a global\nlocation onto each endpoint.\n\nAdd a patch from Rolf fixing 2 issues:\n1. free code did not work right because we were computing the amount we\nfreed after merging the list\n2. we need to store original malloc GPU buffer in extra place because\nthe one in the convertor gets changed over time\n\nbig changes, now pack is driven by receiver by active message\n\nUpon datatype commit create a list of iovec representing a single\niteration of the datatype based on a NULL pointer. This list will\nthen contain the displacement and the length of each fragment of\nthe datatype memory layout and can be used for any packing/unpacking\npurpose.\n\nnow we are able to pack directly to remote buffer if receiver is\ncontiguous\n\nadd ddt_benchmark\n\nreceiver now will send msg back to sender for buffer reuse\n\noffset instead of actual addess, and lots of clean up for unused\nfunctions\n\nre-write pipeline\nrewrite pipeline\n\ns up and running. PUT size in an MCA parameters.\n\nfix pipelining for non-contiguous to contiguous\n\nopal_datatype is chnaged, so we need more space\n\nreorder datatypes to cache boundaries\n\nslience warnings\n\nremove smcuda btl calls from pml ob1\n\nthis file is not used anymore\n\ncuda ddt support is able to turn itself off. Make it support\nmulti-GPU when ompi support multi-GPU in the future\n\nfix a cuda stream bug for iov, remove some stream syncs\n\nin openib, disable rdma for non-contiguous gpu data\n\nmove ddt kernel support function pointer into opal_datatype_cuda.c\n\nrename some functions\n\ncheck point\n\nsupport caching datatype\nAdd support for caching the unpacked datatype description\nvia the opal_convertor_raw_cached function.\n\ncached iov is working for count = 1\ncheck point use raw_cached, but cuda iov caching is not enabled\n\ncheck point, split iov into two version, non-cached and cached\n\ncheck point iov cache\n\nanother checkpoint\n\ncheck point, cuda iov is cached, but not used for pack/unpack\n\ncheck point, ready to use cached cuda iov\n\ncheckpoint, cached cuda iov is working with multiple send, but not for\ncount > 1\n\ncheckpoint, fix a bug for partial unpack\n\ncheckpoint, fix unpack size\n\ncache the entire cuda iov\ncheckpoint, during unpack, cache the entire iov before unpack\n\nanother checkpoint\n\ncheckpoint , remove unnecessary cuda stream sync\n\nuse bit to replace %\n\nrollback to use %, not bit, since it is faster, not sure why\n\nnow cuda iov is {nc_disp, c_disp}\n\nclean up kernel, put variables uses multiple times into register\n\ncached cuda iov is working for count > 1\n\nanother checkpoint\n\nnow convertor->count > 1 is woring\n\nmove the cuda iov caching into a seperate function\n\nthese two variables are useless now\n\nfix a bug for ib, current count of convertor should be set in\nset_cuda_iov_position\n\ncleanup, move cudamalloc into cache cuda iov\n\nrearrange varibles\n\nif cuda_iov is not big enough, use realloc. However, cudaMallocHost does\nnot work with realloc, so use malloc instead\n\nmake sure check pointer is not NULL before free it\n\nrewrite non cached iov, make it unified with cached iov\n\ncheckpoint, rewrite non-cached version\n\nfix for non cached iov\n\nfix the non cached iov, set position should be put at first\n\nmove ddt iov to cuda iov into a function\n\nmerge iov cached and non-cached\n\nfor non cached iov, if there is no enough cuda iov space, break\n\napply loop unroll for pack and unpack kernels\n\napply loop unroll on packing kernels\n\napply unroll to unpack\n\nfix a cuda event bug. cudaStreamWaitEvent is not blocking call.\nfix cuda stream\n\nnew vector kernel\n\nfix a if CUDA_41 error\n\nclean up a if\n\nuse cuda event to track the completion of pack and unpack\nnew for event\n\nevent should work now, but alloc event is not fixed\n\nasync pack unpack for openib\n\ncheckpoint async pack for openib\n\nasync unpack for openib\n\nuse cuda callback to trigger pack and unpack\n\nget rid of cuda stream callback, ompi is not thread safe. Use local cuda event\n\nchange set stream function\n\nmake openib support multi-stream\n\ncreate a btl function to register convertor to registration handle, now we dont need ob1 header in smcuda btl\n\nclean up the pml_ob1_cuda\n\nminor cleanup\n\nfix a bug: we should also track the completion of unpack operation, but only track D2D.\n\ncheckpoint, cached iov now uses seperated block\n\nanother checkpoint\n\nnow non-cached shuuport async operations\n\nuse multiple cuda stream for P2P, it allows multiple send/recv working simutaniously\n\nfix the renaming issue after merge\n\nput ompi_datatype_t back to 512 byte, clean up printf and unused functions\nput ompi_datatype_t back to 512 byte, plus some cleanup\n\nclean up printf, now use OPAL_OUTPUT_VERBOSE\n\nrename function names to opal_datatype_cuda_xxx\n\nmore cleanup\n\nclean up unused function\n\nrevert the ddt_test, will have a separate cuda test later\n\nset the default of ddt pipeline size to 4M\n\nbug fix, set gpu buffer to NULL when init\n\nfix configuration and slient warnings of datatype padding\n\nminor fix in makefile\n\ndisable ddt cuda test\n\nremove unused functions\n\nadd comment for opal_datatype_cuda.cuh, cached device id, remove commented code\n\nmerge pack unpack put sig function into one\n\nrecheck the opal_datatype_cuda_kernel_support\n\nOPAL_DATATYPE_IOV_UNIFIED_MEM is no longer needed\n\nclean up cu files\n\ndisable cuda ddt test to pass make check","shortMessageHtmlLink":"Add GPU packing and unpacking"}},{"before":"73d500a3c9dad5a599abab2591a65c97e9fdefe3","after":"6f31034d2d042309b8081bf243e3d3b00d2e2b31","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-05T22:34:55.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"disable cuda ddt test to pass make check","shortMessageHtmlLink":"disable cuda ddt test to pass make check"}},{"before":"254c5d063e590e5b5499f294b3b00d7470277422","after":"73d500a3c9dad5a599abab2591a65c97e9fdefe3","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-05T22:33:07.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"disable cuda ddt test to pass make check","shortMessageHtmlLink":"disable cuda ddt test to pass make check"}},{"before":null,"after":"254c5d063e590e5b5499f294b3b00d7470277422","ref":"refs/heads/topic/cuda-datatype","pushedAt":"2024-07-05T22:28:34.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"disable cuda ddt test to pass make check","shortMessageHtmlLink":"disable cuda ddt test to pass make check"}},{"before":"5c2dedb861b7db220f89b2546af484885c1493f6","after":null,"ref":"refs/heads/nbuap/topic/dpu_offload","pushedAt":"2024-07-03T17:13:12.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"}},{"before":"17d38734926155b9157f476caab2506d43ac36a6","after":null,"ref":"refs/heads/topic/cuda_ipc_restore","pushedAt":"2024-07-03T17:12:38.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"}},{"before":null,"after":"c6cab81d8e4be48b7d995ae8d0f5f1de15c49c3f","ref":"refs/heads/import/nvidia/ucc/blocking","pushedAt":"2024-07-03T17:09:08.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>\n(cherry picked from commit 5e9ed86eab1a3d14488c1d31995f229c6a6a359d)","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"c29f239e39d5268a3fc454578caeeef8ac841ce1","after":"ff12b69cdd4b09871accd7ac7cd6870f0a55ecb3","ref":"refs/heads/main","pushedAt":"2024-07-03T17:06:25.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12616 from bosilca/nvidia/ucc/blocking\n\nDifferentiate blocking unblocking collectives in UCC","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2351215067\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12616\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12616/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12616\">open-mpi#12616</a> from bosilca/nvidia/ucc/blocking"}},{"before":"5e9ed86eab1a3d14488c1d31995f229c6a6a359d","after":null,"ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-07-03T17:06:02.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"}},{"before":"97e451917403c7b504a7608821e83571ab283d93","after":"5e9ed86eab1a3d14488c1d31995f229c6a6a359d","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-07-02T17:39:40.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"8b7e577c1d4ae97b341a534d881c32535fd629e1","after":"c29f239e39d5268a3fc454578caeeef8ac841ce1","ref":"refs/heads/main","pushedAt":"2024-07-02T16:58:24.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12651 from devreal/workflows-node20\n\nBump workflow actions to node20","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2386246459\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12651\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12651/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12651\">open-mpi#12651</a> from devreal/workflows-node20"}},{"before":"7daa75480f9431d1b4bffb2ca53af6d58429fc85","after":"97e451917403c7b504a7608821e83571ab283d93","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-07-01T19:10:53.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"4e97fb580c5d54c1630d8fdb3d0ae10092bbf4e9","after":"7daa75480f9431d1b4bffb2ca53af6d58429fc85","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-07-01T19:09:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"81ed1492a0f937ca54c0c440a447d3eec99c3452","after":"4e97fb580c5d54c1630d8fdb3d0ae10092bbf4e9","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-07-01T19:05:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"5e45a9dbfdeb140b328864e2d493539ddd0ec78d","after":"8b7e577c1d4ae97b341a534d881c32535fd629e1","ref":"refs/heads/main","pushedAt":"2024-07-01T15:36:15.000Z","pushType":"push","commitsCount":4,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12632 from ggouaillardet/topic/opal_basename\n\nopal: fix opal_basename() for single character filenames","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2361477867\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12632\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12632/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12632\">open-mpi#12632</a> from ggouaillardet/topic/opal_basename"}},{"before":"2aef545cc7b8026299704e5a5c52814827d89bbd","after":"5e45a9dbfdeb140b328864e2d493539ddd0ec78d","ref":"refs/heads/main","pushedAt":"2024-06-20T15:41:10.000Z","pushType":"push","commitsCount":15,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge pull request #12615 from ggouaillardet/topic/op_memory_leaks\n\nop: plug memory leaks","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2350337915\" data-permission-text=\"Title is private\" data-url=\"https://github.com/open-mpi/ompi/issues/12615\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/open-mpi/ompi/pull/12615/hovercard\" href=\"https://github.com/open-mpi/ompi/pull/12615\">open-mpi#12615</a> from ggouaillardet/topic/op_memory_…"}},{"before":"e803b7e4a34b6922cce24c1195d2606af14db17e","after":"abfe91157a3ddd075f2cc62c25688f1310c57fb3","ref":"refs/heads/fix/mpi_dynamic","pushedAt":"2024-06-20T15:38:05.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Trying to fix the connect/accept DPM code.\n\nI still don't think this code is correct, but somehow it works. My main\nconcern is that the PML add_proc will be called with a different list of\nprocesses, or at least in a different order, everywhere. Before this\npatch it worked because we were calling it with ALWAYS 0 procs (because\nby the time we got out of the loop the ilist was empty), so it was not\ndoing much. But, why was it working without a proper add_proc ? Do we\nstill need that add_proc there ?\n\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Trying to fix the connect/accept DPM code."}},{"before":"d62253c90c6095aac01a75b02d260202460799df","after":"81ed1492a0f937ca54c0c440a447d3eec99c3452","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-06-20T15:35:34.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Differentiate blocking unblocking collectives in UCC\n\nSigned-off-by: Geoffroy Vallee <geoffroy.vallee@gmail.com>\nSigned-off-by: George Bosilca <gbosilca@nvidia.com>","shortMessageHtmlLink":"Differentiate blocking unblocking collectives in UCC"}},{"before":"5c2dedb861b7db220f89b2546af484885c1493f6","after":"d62253c90c6095aac01a75b02d260202460799df","ref":"refs/heads/nvidia/ucc/blocking","pushedAt":"2024-06-20T15:33:49.000Z","pushType":"push","commitsCount":11,"pusher":{"login":"bosilca","name":null,"path":"/bosilca","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/642701?s=80&v=4"},"commit":{"message":"Merge branch 'main' into nvidia/ucc/blocking","shortMessageHtmlLink":"Merge branch 'main' into nvidia/ucc/blocking"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0xOFQyMzozNjo0NS4wMDAwMDBazwAAAAS6QLoB","startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0xOFQyMzozNjo0NS4wMDAwMDBazwAAAAS6QLoB","endCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wNi0yMFQxNTozMzo0OS4wMDAwMDBazwAAAARqu3Sz"}},"title":"Activity · bosilca/ompi"}