* Implements --label option for python profiler
- removes '-a', '-f', and '-l' options in favor of '--label args file line' for consistency with omnitrace exe
* cmake formatting
* Bump version to 1.1.0
[ROCm/rocprofiler-systems commit: 017e794d63]
* Implements --label option for python profiler
- removes '-a', '-f', and '-l' options in favor of '--label args file line' for consistency with omnitrace exe
* cmake formatting
* Bump version to 1.1.0
This is a single commit of the source code changes required to
introduce support for multiple ranks per device.
A new interface (ncclCommRankInitMulti) has to be used to make use of
this new feature.
[ROCm/rccl commit: 0336ffdf70]
This is a single commit of the source code changes required to
introduce support for multiple ranks per device.
A new interface (ncclCommRankInitMulti) has to be used to make use of
this new feature.
* Fix sampling counter time scales
- All perfetto trace events have "begin_ns" and "end_ns" debug fields
- data for thread start and end timestamp in pthread_create_gotcha
- discard samples outside of thread start and end timestamps
- rename "CPU User CPU Time" perfetto counter to "CPU User Time"
- rename "CPU Kernel CPU Time" perfetto counter to "CPU Kernel Time"
- ensure CPU system samples in perfetto are set to zero at end
- backtrace uses comp::wall_clock record() for timestamps (consistency)
- "Peak Memory Usage [Thread X] (S)" renamed to "Thread Peak Memory Usage [X] (S)"
- "Context Switches [Thread X] (S)" renamed to "Thread Context Switches Usage [X] (S)"
- "Page Faults [Thread X] (S)" renamed to "Thread Page Faults Usage [X] (S)"
- "<PAPI_DESC> [Thread X] (S)" renamed to "Thread <PAPI_DESC> [X] (S)"
- samples
* Fix includes
[ROCm/rocprofiler-systems commit: 90ab7a89fc]
* Fix sampling counter time scales
- All perfetto trace events have "begin_ns" and "end_ns" debug fields
- data for thread start and end timestamp in pthread_create_gotcha
- discard samples outside of thread start and end timestamps
- rename "CPU User CPU Time" perfetto counter to "CPU User Time"
- rename "CPU Kernel CPU Time" perfetto counter to "CPU Kernel Time"
- ensure CPU system samples in perfetto are set to zero at end
- backtrace uses comp::wall_clock record() for timestamps (consistency)
- "Peak Memory Usage [Thread X] (S)" renamed to "Thread Peak Memory Usage [X] (S)"
- "Context Switches [Thread X] (S)" renamed to "Thread Context Switches Usage [X] (S)"
- "Page Faults [Thread X] (S)" renamed to "Thread Page Faults Usage [X] (S)"
- "<PAPI_DESC> [Thread X] (S)" renamed to "Thread <PAPI_DESC> [X] (S)"
- samples
* Fix includes
- fix loop-level instrumentation
- support loop instrumentation w/o debug symbols via loop number
- improve module_function messages
- serialize num_basic_blocks
- serialize num_outer_loops
- serialize is_num_instructions_constrained
- serialize is_loop_num_instructions_constrained
- updated transpose example to use uniform_int_distribution
- added transpose loop test
- added fail regexes for tests which enable loop instrumentation
- use module->getFullName in get_loop_file_line_info
- use module->getFullName in get_func_file_line_info
- use module->getFullName in get_basic_block_file_line_info
[ROCm/rocprofiler-systems commit: a640fbdb29]
- fix loop-level instrumentation
- support loop instrumentation w/o debug symbols via loop number
- improve module_function messages
- serialize num_basic_blocks
- serialize num_outer_loops
- serialize is_num_instructions_constrained
- serialize is_loop_num_instructions_constrained
- updated transpose example to use uniform_int_distribution
- added transpose loop test
- added fail regexes for tests which enable loop instrumentation
- use module->getFullName in get_loop_file_line_info
- use module->getFullName in get_func_file_line_info
- use module->getFullName in get_basic_block_file_line_info
Windows complains about undefined symbol for
hipStreamWrite32/64 and hipStreamWait32/64 -
commit hash 08435a972b
Change-Id: I48b9a6591cab3a05baaedbc26030914dd930fb68
[ROCm/clr commit: de026633c0]
Windows complains about undefined symbol for
hipStreamWrite32/64 and hipStreamWait32/64 -
commit hash 08435a972b
Change-Id: I48b9a6591cab3a05baaedbc26030914dd930fb68
The same information can be generated from the hcc_ops_trace.txt file,
so in a later commit, will add a stage to the tblextr.py script to
generate the .csv files when ROCP_STATS_OPT=1.
Change-Id: I3d1575e096bedf98c66068d9a4ca141421e5bb9d
[ROCm/roctracer commit: 836bab37d3]
The same information can be generated from the hcc_ops_trace.txt file,
so in a later commit, will add a stage to the tblextr.py script to
generate the .csv files when ROCP_STATS_OPT=1.
Change-Id: I3d1575e096bedf98c66068d9a4ca141421e5bb9d
* Rework sampling trace counter names + new trace counters
- reformulate trace counter names for easier comparison (thread sampling)
- new process-level trace counters for context switches (thread sampling)
- new process-level trace counters for page faults (thread sampling)
- new process-level trace counters for CPU time (thread sampling)
- new thread-level trace counters for context switches (sampling)
- new thread-level trace counters for page faults (sampling)
* tweak header include in backtrace.cpp
[ROCm/rocprofiler-systems commit: 1db3934f85]
* Rework sampling trace counter names + new trace counters
- reformulate trace counter names for easier comparison (thread sampling)
- new process-level trace counters for context switches (thread sampling)
- new process-level trace counters for page faults (thread sampling)
- new process-level trace counters for CPU time (thread sampling)
- new thread-level trace counters for context switches (sampling)
- new thread-level trace counters for page faults (sampling)
* tweak header include in backtrace.cpp
Fixes for hipGraphInstantiate OutofMemory issue and
soft hang of hipStreamSynchronize.
Change-Id: Ic2bb97582c5978ef2e4eb3a1b124483aa412838e
[ROCm/clr commit: c06bfa6bef]
Jatin Chaudhary
Satyanvesh Dittakavi
Jonathan R. Madsen
Edgar
Jenkins
Maneesh Gupta
Hao Zhou
Ajay
Laurent Morichetti
sdashmiz
Bill(Shuzhou) Liu