Implement std::mutex based monitor that has much
simpler logics than legacy monitor.
Create DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR to
toggle them.
If DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR = false
(by default), use legacy monitor;
If DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR = true,
use std::mutex based monitor.
If no perf drop of stl::mutex based monitor,
legacy one will be removed later.
Change-Id: I1d21368ff462477d3238d71e4e2a1a7d6b9167ad
[ROCm/clr commit: 73c02041e1]
Implement std::mutex based monitor that has much
simpler logics than legacy monitor.
Create DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR to
toggle them.
If DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR = false
(by default), use legacy monitor;
If DEBUG_CLR_USE_STDMUTEX_IN_AMD_MONITOR = true,
use std::mutex based monitor.
If no perf drop of stl::mutex based monitor,
legacy one will be removed later.
Change-Id: I1d21368ff462477d3238d71e4e2a1a7d6b9167ad
MSCCL can now run in a multi-threaded configuration. To test in the unit tests, added the ENABLE_OPENMP compile definition flag and the --openmp-test-enable flag to the unit test build script. To activate, set the environment variables UT_MULTITHREADED=1 and UT_PROCESS_MASK=1. Set Jenkins to use this mode.
[ROCm/rccl commit: 0c36d571ea]
MSCCL can now run in a multi-threaded configuration. To test in the unit tests, added the ENABLE_OPENMP compile definition flag and the --openmp-test-enable flag to the unit test build script. To activate, set the environment variables UT_MULTITHREADED=1 and UT_PROCESS_MASK=1. Set Jenkins to use this mode.
* General fixes to ATT, packets and event ID retrieval
* Update source/lib/rocprofiler-sdk/hsa/aql_packet.hpp
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
* PC sampling: integration test with instruction decoding
* PC sampling: verifying internal and external CIDs
The PC sampling integration test has been extended
to verify internal and external correlation IDs.
* tmp solution of using Instructions as keys
* wrapper for HIP call
* PCS integration test: ld_addr as instruction id
For the sake of the integration test, use as the
instruction identifier. To support code object unloading
and relocations, use as the identifier
(the change in the decoder is required).
* PCS integration test: removing shared_ptr
Completely removing usage of shared pointers.
* PCS integration test: removing decoder
When a code object has been unloaded, ensure all PC samples
corresponding to that object are decoded, prior to removing
the decoder.
* PCS integration test: fixing build flags and imports
* PCS integration test: fixing labels
* PCS integration test: cmake flags fix
* PC sampling cmake labels renamed
* PCS integration test refactoring
* PCS integration test: minimize usage of raw pointers
* PCS integration test: at least one sample should be delivered.
* PC sampling lables: pc-sampling
[ROCm/rocprofiler-sdk commit: 498b1f2bd7]
* PC sampling: integration test with instruction decoding
* PC sampling: verifying internal and external CIDs
The PC sampling integration test has been extended
to verify internal and external correlation IDs.
* tmp solution of using Instructions as keys
* wrapper for HIP call
* PCS integration test: ld_addr as instruction id
For the sake of the integration test, use as the
instruction identifier. To support code object unloading
and relocations, use as the identifier
(the change in the decoder is required).
* PCS integration test: removing shared_ptr
Completely removing usage of shared pointers.
* PCS integration test: removing decoder
When a code object has been unloaded, ensure all PC samples
corresponding to that object are decoded, prior to removing
the decoder.
* PCS integration test: fixing build flags and imports
* PCS integration test: fixing labels
* PCS integration test: cmake flags fix
* PC sampling cmake labels renamed
* PCS integration test refactoring
* PCS integration test: minimize usage of raw pointers
* PCS integration test: at least one sample should be delivered.
* PC sampling lables: pc-sampling
- Introduce a lock when checking isUserObjectValid. We need a lock
here as one can remove the userObject T2, leading to buffer overflow
when checking ranges in T1.
Change-Id: I058144b8cc463c90ab6bf5cf96bf937897742917
[ROCm/clr commit: 6ac67afdd5]
- Introduce a lock when checking isUserObjectValid. We need a lock
here as one can remove the userObject T2, leading to buffer overflow
when checking ranges in T1.
Change-Id: I058144b8cc463c90ab6bf5cf96bf937897742917
- omnitrace-run, omnitrace-sample, and omnitrace-causal now automatically append the LD_LIBRARY_PATH with the directory containing the omnitrace libraries
- this helps ensure that binary rewritten exes can resolve omnitrace-rt library location
[ROCm/rocprofiler-systems commit: 18833a0a5e]
- omnitrace-run, omnitrace-sample, and omnitrace-causal now automatically append the LD_LIBRARY_PATH with the directory containing the omnitrace libraries
- this helps ensure that binary rewritten exes can resolve omnitrace-rt library location
Aryan Salmanpour
Giovanni LB
jujiang
Jaydeep Patel
Tao Sang
corey-derochie-amd
Ioannis Assiouras
Zhang Ava
Vladimir Indic
Saleel Kudchadker
Wenkai Du
Chris Freehill
Maneesh Gupta
dependabot[bot]
sdashmiz
Jonathan R. Madsen