Check for RLIMIT_CORE before collecting data for coredump. If the
current limit is 0, then we can return early without spending time
collecting coredump data.
[ROCm/ROCR-Runtime commit: d031af9eb5]
To execute set and reset commands, amd-smi needs sudo
privileges. Updated the subcommand help text to show
'sudo' requirement for these commands.
Signed-off-by: Bindhiya Kanangot Balakrishnan <Bindhiya.KanangotBalakrishnan@amd.com>
Co-authored-by: Arif, Maisam <Maisam.Arif@amd.com>
To execute set and reset commands, amd-smi needs sudo
privileges. Updated the subcommand help text to show
'sudo' requirement for these commands.
Signed-off-by: Bindhiya Kanangot Balakrishnan <Bindhiya.KanangotBalakrishnan@amd.com>
Co-authored-by: Arif, Maisam <Maisam.Arif@amd.com>
[ROCm/amdsmi commit: 2141e0336c]
- The Perfetto validation script was generating false positives if the queried Perfetto data was empty.
- Fixed some failures that appeared after fixing the script.
[ROCm/rocprofiler-systems commit: 610cd00971]
- The Perfetto validation script was generating false positives if the queried Perfetto data was empty.
- Fixed some failures that appeared after fixing the script.
Since hipMemMap can be called for multiple device handles on the same virtual memory, the same is true for hipMemUnmap, meaning that virtual memory can be "partially unmapped".
This means that the unmap function can be called for a specific part of the reserved address, meaning that only the designated subbuffer should be released. If unmap is called on the entire reserved memory, then all subbuffers should be released.
The main point is that for every hsa_amd_vmem_map, there should be a corresponding hsa_amd_vmem_unmap. Otherwise, if entire memory is unmapped by a single unmap call, then HSA will report the memory as "in use" if an attempt is made to delete it.
Change-Id: I039308eafb820decfb1c09f60347f26cdad1a362
[ROCm/clr commit: 3ec1d2d2f1]
Since hipMemMap can be called for multiple device handles on the same virtual memory, the same is true for hipMemUnmap, meaning that virtual memory can be "partially unmapped".
This means that the unmap function can be called for a specific part of the reserved address, meaning that only the designated subbuffer should be released. If unmap is called on the entire reserved memory, then all subbuffers should be released.
The main point is that for every hsa_amd_vmem_map, there should be a corresponding hsa_amd_vmem_unmap. Otherwise, if entire memory is unmapped by a single unmap call, then HSA will report the memory as "in use" if an attempt is made to delete it.
Change-Id: I039308eafb820decfb1c09f60347f26cdad1a362
- Use getBuffer/releaseBuffer in BlitManager
- Cleanup XferBuffer as we use ManagedBuffer for both reads/writes
Change-Id: I2661b85dd012763b17a38a743fec1b1d79125f67
[ROCm/clr commit: 37d606d193]
- Use getBuffer/releaseBuffer in BlitManager
- Cleanup XferBuffer as we use ManagedBuffer for both reads/writes
Change-Id: I2661b85dd012763b17a38a743fec1b1d79125f67
- Add document formatting workflow from ROCm/rocm-docs-core/.github/workflows/linting.yml
- Updated runners to use ubuntu-22.04, since 20.04 is being deprecated.
[ROCm/rocprofiler-systems commit: a18d00d784]
- Add document formatting workflow from ROCm/rocm-docs-core/.github/workflows/linting.yml
- Updated runners to use ubuntu-22.04, since 20.04 is being deprecated.
Modify test case to support XL cards.
Change-Id: I6ad45a290d50a5238804ce7417bcdb33a3912872
Signed-off-by: David Belanger <david.belanger@amd.com>
[ROCm/ROCR-Runtime commit: 3ceb131df5]
- If any kernel uses device heap, the launch needs to be preceeded by an
init kernel, Save on the extra barrier packet launch/flush between the
init heap kernel and user kernel
Change-Id: I8ebc6246188200e5f673dc464bc76a53bcb8b7c6
[ROCm/clr commit: ca530c660b]
- If any kernel uses device heap, the launch needs to be preceeded by an
init kernel, Save on the extra barrier packet launch/flush between the
init heap kernel and user kernel
Change-Id: I8ebc6246188200e5f673dc464bc76a53bcb8b7c6
* Add check to skip counter_storage::write() if internal storage field is destroyed.
* Output warning message if counter data is not available when trying to write out to Timemory
---------
Co-authored-by: David Galiffi <David.Galiffi@amd.com>
[ROCm/rocprofiler-systems commit: 43f900d01e]
David Yat Sin
Arif, Maisam
adapryor
Galantsev, Dmitrii
Li Ma
dependabot[bot]
Bhardwaj, Gopesh
Ioannis Assiouras
Kanangot Balakrishnan, Bindhiya
Jaydeep Patel
David Galiffi
Aidan Belton-Schure
Marko Arandjelovic
Todd tiantuo Li
Saleel Kudchadker
Rawat, Swati
David Belanger
Fei Zheng
Sohaib Nadeem