1. Create P2P links
2. Determine FRAMEBUFFER_PUBLIC/PRIVATE only based
host-accessibility, not peer-accesssibility
Signed-off-by: Gang Ba <gaba@amd.com>
Change-Id: I15fccdc60386b453e2a47849a16df15157324b21
1. Create P2P links
2. Determine FRAMEBUFFER_PUBLIC/PRIVATE only based
host-accessibility, not peer-accesssibility
Signed-off-by: Gang Ba <gaba@amd.com>
Change-Id: I15fccdc60386b453e2a47849a16df15157324b21
[ROCm/ROCR-Runtime commit: bedecc5957]
Optimization for Tree allreduce on A100.
Improve aggregation performance.
Use shared buffers for inter-node send/recv.
Add NVTX profiling hooks.
Accelerate alltoall connections by merging communication for all
channels.
Add support for one hop communication through NVLink, for faster
send/recv communication on cubemesh topologies like DGX-1.
Improve alltoall scheduling to better balance intra/inter node
communication.
Increase send/recv parallelism by 8x, each warp sending or
receiving to a different peer.
Net: move to v4.
Net: make flush operation asynchronous to accelerate alltoall.
Net: define maximum number of requests.
Fix hang when using LL128 protocol after 2^31 steps.
Fix#379 : topology injection failing when using less GPUs than
described in the XML.
Fix#394 : protocol mismatch causing hangs or crashes when using
one GPU per node.
[ROCm/rccl commit: 920dbe5b35]
Optimization for Tree allreduce on A100.
Improve aggregation performance.
Use shared buffers for inter-node send/recv.
Add NVTX profiling hooks.
Accelerate alltoall connections by merging communication for all
channels.
Add support for one hop communication through NVLink, for faster
send/recv communication on cubemesh topologies like DGX-1.
Improve alltoall scheduling to better balance intra/inter node
communication.
Increase send/recv parallelism by 8x, each warp sending or
receiving to a different peer.
Net: move to v4.
Net: make flush operation asynchronous to accelerate alltoall.
Net: define maximum number of requests.
Fix hang when using LL128 protocol after 2^31 steps.
Fix#379 : topology injection failing when using less GPUs than
described in the XML.
Fix#394 : protocol mismatch causing hangs or crashes when using
one GPU per node.
RPM needs _REQUIRES at the end, not _DEPENDS, and also requires a space
before the version of the required package.
Change-Id: I9dd70bd92fc2407b7e8b31e4d46df43c52438a65
Signed-off-by: Kent Russell <kent.russell@amd.com>
RPM needs _REQUIRES at the end, not _DEPENDS, and also requires a space
before the version of the required package.
Change-Id: I9dd70bd92fc2407b7e8b31e4d46df43c52438a65
Signed-off-by: Kent Russell <kent.russell@amd.com>
[ROCm/ROCR-Runtime commit: 089fdeb1fe]
* gtest: add scatter to combined calls and use loops
* gtest: run validation inside loop
* gtest: revert small element count to 2520
* gtest: fix memory leak in validation
[ROCm/rccl commit: b0853ccd51]
* gtest: add scatter to combined calls and use loops
* gtest: run validation inside loop
* gtest: revert small element count to 2520
* gtest: fix memory leak in validation
Remove the following from hip cmakefile,
1.find_package cmds of hsa-runtime64 and amd_comgr;
2.macros of USE_COMGR_LIBRARY and COMGR_DYN_DLL
because they have already been used in Rocclr cmakefile.
cmake mechanism will make hip cmakefile inherite them.
Change-Id: I03b527fc4135c79bb35cbbfe419211ec363cb067
[ROCm/clr commit: 10f4918d97]
Remove the following from hip cmakefile,
1.find_package cmds of hsa-runtime64 and amd_comgr;
2.macros of USE_COMGR_LIBRARY and COMGR_DYN_DLL
because they have already been used in Rocclr cmakefile.
cmake mechanism will make hip cmakefile inherite them.
Change-Id: I03b527fc4135c79bb35cbbfe419211ec363cb067
Remove the following from hip cmakefile,
1.find_package cmds of hsa-runtime64 and amd_comgr;
2.macros of USE_COMGR_LIBRARY and COMGR_DYN_DLL
because they have already been used in Rocclr cmakefile.
cmake mechanism will make hip cmakefile inherite them.
Change-Id: I03b527fc4135c79bb35cbbfe419211ec363cb067
[ROCm/hip commit: ecd4dca29d]
Remove the following from hip cmakefile,
1.find_package cmds of hsa-runtime64 and amd_comgr;
2.macros of USE_COMGR_LIBRARY and COMGR_DYN_DLL
because they have already been used in Rocclr cmakefile.
cmake mechanism will make hip cmakefile inherite them.
Change-Id: I03b527fc4135c79bb35cbbfe419211ec363cb067
The hack dosn't really track the commands status. It may be not
necessary for HIP, but will cause early resource release.
Change-Id: I791ad36dd8abd3b6b3d2c9b16a210a555c08ca64
[ROCm/clr commit: 532f0ae951]
The hack dosn't really track the commands status. It may be not
necessary for HIP, but will cause early resource release.
Change-Id: I791ad36dd8abd3b6b3d2c9b16a210a555c08ca64
Wenkai Du
Gang Ba
Michael LIAO
Sylvain Jeaugey
Vladislav Sytchenko
Kent Russell
Rahul Garg
Wenkai Du
Chauncey Hui
Tao Sang
Stanley Tsang
gilbertlee-amd
German Andryeyev
Jason Tang
Sarbojit Sarkar