* rocprofv3: suppress agent info when no data collected
* Update output config serialization
- full serialization of output configuration
* Update rocprofiler-sdk-att/tests
- add version and soversion
- change output directory
- generate libatt_decoder_summary
- disable tests instead of removing them
* Update rocprofv3 command-line
- make --att-library-path hidden by default
- simplify check_att_capability
- reorder pc sampling options
- add hidden --echo option
- remove ROCPROF_LIST_AVAIL_TOOL_LIBRARY from preload
* Add new rocprofv3 tests for specify the ATT library path
* Tweak to rocprofv3-test-hsa-multiqueue-att tests
* Update rocprofv3 tool to enable output with att
* Fix standalone test installation
* Revert to fetchcontent_makeavailable to fetchcontent_populate
* Revert tests/common/CMakeLists.txt
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 59b41ab5aa]
* [DO NOT MERGE] Misc UUID updates
- this is WIP
* Agent visibility
- Support for ROCR_VISIBLE_DEVICES, HIP_VISIBLE_DEVICES, CUDA_VISIBLE_DEVICES, GPU_DEVICE_ORDINAL
* Update CHANGELOG
* tweak to rocprofiler_agent_runtime_visiblity_t
* Code object kernel address
- new fields in code_object_kernel_symbol_register_data_t
- kernel_code_entry_byte_offset
- kernel_address
* Support ROCR_VISIBLE_DEVICES reordering devices for HIP
* Addressed code review changes
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 6246ec4040]
* rocprofv3: do not abort if counter does not have dimensions
* Relax error handling further in rocprofv3 metadata
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 3071199386]
* Force HSA_AMD_MEMORY_POOL_EXECUTABLE_FLAG value to be used with HSA calls
Fix for CI
* More tweaks
* Increase reproducible-runtime kernel sleep granularity
* Fix data race in synchronous device counter collection sample
* Update device counting service
- add get_active_context function
---------
Co-authored-by: Benjamin Welton <bewelton@amd.com>
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 080b2ba451]
* Add regex for undefined behavior to ROCPROFILER_DEFAULT_FAIL_REGEX
- add UBSAN_OPTIONS to setup-sanitizer-env.sh
* Improve ROCPROFILER_DEFAULT_FAIL_REGEX
* Use -fno-sanitize-recover=undefined flag
- this compiler flag causes all undefined behavior errors to exit
* Revert ROCPROFILER_DEFAULT_FAIL_REGEX
* fix for shift overflow
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
Co-authored-by: Manjunath-Jakaraddi <manjunath.jakaraddi@amd.com>
[ROCm/rocprofiler-sdk commit: e743bf5a93]
* SDK: No bg thread if no clients use SDK
* Update CHANGELOG
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 0fbe6cc7b6]
* Adding support for hsa_amd_signal_wait_all
* Fixes for HIP
---------
Co-authored-by: Ammar ELWazir <aelwazir@amd.com>
[ROCm/rocprofiler-sdk commit: 02a519e84e]
* Adding New HIP APIs
* Format Fix
* Format Fix
* Removing changes from ostream and moving it to format
* Addressing Code Review Comments
* Versioning the new hip calls formatting
---------
Co-authored-by: Ammar ELWazir <aelwazir@amd.com>
[ROCm/rocprofiler-sdk commit: dd5c0ea257]
* [SWDEV-509876] Remove buffer requirement from device counting service
No longer require a buffer to be given when setting up device counting
service. This is to reduce performance overhead in cases where immediate
return of counting samples is being used (synchronous mode).
* Missed file
* Update source/include/rocprofiler-sdk/device_counting_service.h
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
* Update source/lib/rocprofiler-sdk/counters/controller.cpp
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
* Update source/lib/rocprofiler-sdk/counters/device_counting.cpp
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
* Fixes for build
---------
Co-authored-by: Benjamin Welton <ben@amd.com>
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
Co-authored-by: Benjamin Welton <bewelton@amd.com>
[ROCm/rocprofiler-sdk commit: 0c4a56c6bb]
Fix HSA_AMD_MEMORY_POOL_EXECUTABLE_FLAG for ROCm < 6.4
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 72a27feb04]
* Fix async copy validation test
- make the async copy tracing test work regardless of however many HSA memory copies the HIP memory copy decomposes into
* Fix rocprofv3 memory copy tests
* Fix compilation support for hipGraphBatchMemOpNodeGetParams
* Fix rocprofv3-test-summary-*-validate
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: 7fcd80f744]
Fix HIP data type stringify
- when ROCPROFILER_CI is not defined, provide default for case statements
- Add support for hipGraphNodeTypeBatchMemOp when HIP version is >= 6.4.0
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
[ROCm/rocprofiler-sdk commit: f3752faa0a]
We need execute permission for HSA memory (req for IB buffers).
Enforcement is upcoming which will break counter collection (see
ticket).
Co-authored-by: Benjamin Welton <bewelton@amd.com>
Co-authored-by: Bhardwaj, Gopesh <Gopesh.Bhardwaj@amd.com>
[ROCm/rocprofiler-sdk commit: 0d701cdaac]
* rocDecode API Tracing support
* Test bin file added to rocdecode. Need to add validate python methods
* Added option to not make rocDecode tests
* Added rocdecode and rocprofv3 tests
* Added csv test
* Address PR comments. Changed tests to use built-in rocstreambit decoder to remove ffmpeg dependancy. Changed cmake option to disbale tests rather than not build them. Tests work locally, but will fail until rocDecode is built with tracing enabled on CI
* Add option to avoid building rocdecode tests
* Added option to avoid building rocdecode bin file
* Merge conflict error
* CMake files changed in response to review comments. Attempting to implement callbacks.
* Turned off test building for rocdecode
* Minor fixes for review comments
* Review comments
* Updated formatting
* Document changes and format.hpp reversion. Need to remove iterate args support for now for later update.
* Remove iterate args support
* Remove iterate-args
* enforce abi versioning in macro if
* Fix doc error
* removed spaces to fix indentation error
---------
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
[ROCm/rocprofiler-sdk commit: e307b89ca4]
* [SWDEV-509659] Skip rocprof device counting tests if lacking permissions
Skips non-intercept test if proper permissions are not obtained
(SYS_PERFMON). This should be the only test that fails due to permission
issues (others do not require the IOCTL to pass).
Regex match sample: https://regexr.com/8b29s
* Update source/lib/rocprofiler-sdk/counters/tests/CMakeLists.txt
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
* fix
* Update source/lib/rocprofiler-sdk/counters/tests/CMakeLists.txt
---------
Co-authored-by: Benjamin Welton <ben@amd.com>
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
[ROCm/rocprofiler-sdk commit: 536fbba627]
* Split ABI checks for rocprofiler-sdk-roctx into separate file
* Update source/lib/rocprofiler-sdk-roctx/abi.cpp
Co-authored-by: Indic, Vladimir <Vladimir.Indic@amd.com>
* Update source/lib/rocprofiler-sdk-roctx/abi.cpp
Co-authored-by: Indic, Vladimir <Vladimir.Indic@amd.com>
* New line
---------
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
Co-authored-by: Indic, Vladimir <Vladimir.Indic@amd.com>
[ROCm/rocprofiler-sdk commit: 1b7ab08ded]
* Adding source snapshot
* Adding option to serialize only on target kernel
* Fix for tidy
* Formatting
* Testing the new flag
---------
Co-authored-by: Giovanni Baraldi <gbaraldi@amd.com>
[ROCm/rocprofiler-sdk commit: a2fa188e14]
If a user requests PC sampling on a system that does not support this feature,
report a fatal error message and stop executing the program.
[ROCm/rocprofiler-sdk commit: 0ce75c1043]
Add support for device counter colleciton ioctl
Adds support for the device counter collection IOCTL. This IOCTL
allows for device wide counters to be collected even if the queue
is not intercepted by rocprofiler-sdk (required for system profilers).
A test is also included which checks this behavior by creating a queue
that does not have profiling enabled on it and checks to see if SQ
counters can be read from it. Note: this test will be skipped if the KFD
version does not contain this IOCTL.
Right now the check is "soft" in that if the IOCTL is present and there
is an error with permissions, rocprofiler will continue but will print
an error stating that system wide device profiling and collected counter
values may be degraded. This is primarily to avoid breaking existing
users (like PAPI) who may not need the IOCTL's capability and to give
them time to update.
Co-authored-by: Benjamin Welton <ben@amd.com>
[ROCm/rocprofiler-sdk commit: c574881cdb]
* PC Sampling API: emit info logs instead of error
Inside PC sampling API, emit info logs instead of
error logs. The tests verifies status code of each
API call and decide when to skip, instead of relying
on messages in logs.
The samples_processing.cpp test has been removed as it's
not used.
[ROCm/rocprofiler-sdk commit: b4d7ee7887]
Madsen, Jonathan
Kandula, Venkateshwar reddy
Welton, Benjamin
Indic, Vladimir
Elwazir, Ammar
U, Srihari
Nagaraj, Sriraksha
Jakaraddi, Manjunath
Kuricheti, Mythreya
Baraldi, Giovanni
Rawat, Swati
Trowbridge, Ian