[rocpd] Update CSV to SDK method of writing CSVs

.github/workflows/code_coverage.yml

@@ -56,11 +56,11 @@ jobs:
     strategy:
       # fail-fast: false
       matrix:
-        runner: ['rocprof-azure']
+        runner: ['mi300x']
         os: ['ubuntu-22.04']
         build-type: ['Release']
 
-    runs-on: ${{ matrix.runner }}-emu-runner-set
+    runs-on: rocprof-ubuntu-22
 
     # define this for containers
     env:

.github/workflows/continuous_integration.yml

@@ -52,12 +52,13 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        runner: ['navi3', 'navi4', 'mi3xx']
-        os: ['ubuntu-22.04']
-        build-type: ['RelWithDebInfo']
-        ci-flags: ['--linter clang-tidy']
+        system: [
+          {gpu: 'navi3', runner: 'navi3-emu-runner-set', os: 'ubuntu-22.04', build-type: 'RelWithDebInfo', ci-flags: '--linter clang-tidy'},
+          {gpu: 'navi4', runner: 'navi4-emu-runner-set', os: 'ubuntu-22.04', build-type: 'RelWithDebInfo', ci-flags: '--linter clang-tidy'},
+          {gpu: 'mi3xx', runner: 'rocprof-ubuntu-22', os: 'ubuntu-22.04', build-type: 'RelWithDebInfo', ci-flags: '--linter clang-tidy'}
+        ]
 
-    runs-on: ${{ matrix.runner == 'mi3xx' && 'rocprof-azure' || matrix.runner }}-emu-runner-set
+    runs-on: ${{ matrix.system.runner }}
 
     # define this for containers
     env:
@@ -90,7 +91,7 @@ jobs:
         ls -la
 
     - name: Enable PC Sampling
-      if: ${{ contains(matrix.runner, 'mi200') || contains(matrix.runner, 'mi300a') }}
+      if: ${{ contains(matrix.system.gpu, 'mi200') || contains(matrix.system.gpu, 'mi300a') }}
       shell: bash
       run: |
           echo 'ROCPROFILER_PC_SAMPLING_BETA_ENABLED=1' >> $GITHUB_ENV
@@ -100,39 +101,39 @@ jobs:
       shell: bash
       run:
         python3 ./source/scripts/run-ci.py -B build
-          --name ${{ github.repository }}-${{ github.ref_name }}-${{ matrix.os }}-${{ matrix.runner }}-core
+          --name ${{ github.repository }}-${{ github.ref_name }}-${{ matrix.system.os }}-${{ matrix.system.gpu }}-core
           --build-jobs 16
           --site ${RUNNER_HOSTNAME}
           --gpu-targets ${{ env.GPU_TARGETS }}
           --run-attempt ${{ github.run_attempt }}
-          ${{ matrix.ci-flags }}
+          ${{ matrix.system.ci-flags }}
           --
           -DROCPROFILER_DEP_ROCMCORE=ON
           -DROCPROFILER_BUILD_DOCS=OFF
-          -DCMAKE_BUILD_TYPE=${{ matrix.build-type }}
+          -DCMAKE_BUILD_TYPE=${{ matrix.system.build-type }}
           -DCMAKE_INSTALL_PREFIX=/opt/rocprofiler-sdk
           -DCPACK_GENERATOR='DEB;RPM;TGZ'
           -DCPACK_PACKAGING_INSTALL_PREFIX="$(realpath /opt/rocm)"
           -DPython3_EXECUTABLE=$(which python3)
           ${{ env.GLOBAL_CMAKE_OPTIONS }}
           --
-          -LE "${${{ matrix.runner }}_EXCLUDE_LABEL_REGEX}"
-          -E "${${{ matrix.runner }}_EXCLUDE_TESTS_REGEX}"
+          -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}"
+          -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}"
 
     - name: Install
-      if: ${{ contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       timeout-minutes: 10
       run:
         cmake --build build --target install --parallel 16
 
     - name: Build Packaging
-      if: ${{ contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       timeout-minutes: 10
       run:
         cmake --build build --target package --parallel 16
 
     - name: Test Install Build
-      if: ${{ contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       timeout-minutes: 20
       shell: bash
       run: |
@@ -141,11 +142,11 @@ jobs:
         export LD_LIBRARY_PATH=/opt/rocprofiler-sdk/lib:${LD_LIBRARY_PATH}
         cmake --build build-samples --target all --parallel 16
         cmake --build build-tests --target all --parallel 16
-        ctest --test-dir build-samples -LE "${${{ matrix.runner }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.runner }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
-        ctest --test-dir build-tests -LE "${${{ matrix.runner }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.runner }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
+        ctest --test-dir build-samples -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
+        ctest --test-dir build-tests -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
 
     - name: Install Packages
-      if: ${{ contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       timeout-minutes: 5
       shell: bash
       run: |
@@ -157,19 +158,19 @@ jobs:
         for i in $(ls -S ./build/rocprofiler-sdk*.deb | egrep -v 'roctx|rocpd'); do dpkg --force-all -i ${i}; done;
 
     - name: Test Installed Packages
-      if: ${{ contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       timeout-minutes: 20
       shell: bash
       run: |
         CMAKE_PREFIX_PATH=/opt/rocm cmake -B build-samples-deb /opt/rocm/share/rocprofiler-sdk/samples
         CMAKE_PREFIX_PATH=/opt/rocm cmake -B build-tests-deb -DGPU_TARGETS="gfx942" /opt/rocm/share/rocprofiler-sdk/tests
         cmake --build build-samples-deb --target all --parallel 16
         cmake --build build-tests-deb --target all --parallel 16
-        ctest --test-dir build-samples-deb -LE "${${{ matrix.runner }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.runner }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
-        ctest --test-dir build-tests-deb -LE "${${{ matrix.runner }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.runner }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
+        ctest --test-dir build-samples-deb -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
+        ctest --test-dir build-tests-deb -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}" -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}" --output-on-failure
 
     - name: Archive production artifacts
-      if: ${{ github.event_name == 'workflow_dispatch' && contains(matrix.runner, env.CORE_EXT_RUNNER) }}
+      if: ${{ github.event_name == 'workflow_dispatch' && contains(matrix.system.gpu, env.CORE_EXT_RUNNER) }}
       uses: actions/upload-artifact@v4
       with:
         name: installers-deb
@@ -184,11 +185,11 @@ jobs:
       fail-fast: false
       matrix:
         runner: ['mi300']
-        os: ['rhel', 'sles']
+        os: ['rhel-8', 'sles-15']
         build-type: ['RelWithDebInfo']
         ci-flags: ['']
 
-    runs-on: rocprof-azure-${{ matrix.os }}-emu-runner-set
+    runs-on: rocprof-${{ matrix.os }}
 
     # define this for containers
     env:
@@ -203,7 +204,7 @@ jobs:
       shell: bash
       run: |
         git config --global --add safe.directory '*'
-        if [ "${OS_TYPE}" == "rhel" ]; then
+        if [ "${OS_TYPE}" == "rhel-8" ]; then
           dnf makecache
           dnf groupinstall -y "Development Tools"
           dnf remove -y gcc-c++
@@ -261,7 +262,7 @@ jobs:
         build-type: ['RelWithDebInfo']
 
     if: ${{ contains(github.event_name, 'pull_request') }}
-    runs-on: ${{ matrix.runner == 'mi3xx' && 'rocprof-azure' || matrix.runner }}-emu-runner-set
+    runs-on: ${{ matrix.runner == 'mi3xx' && 'rocprof-ubuntu-22' || matrix.runner }}
 
     # define this for containers
     env:

CHANGELOG.md

@@ -174,7 +174,7 @@ Full documentation for ROCprofiler-SDK is available at [rocm.docs.amd.com/projec
   - absolute == node_id
   - relative == logical_node_id
   - type-relative == logical_node_type_id
-- Added MI300 stochastic (hardware-based) PC sampling support in ROCProfiler-SDK and ROCProfV3
+- Added MI300/MI350 stochastic (hardware-based) PC sampling support in ROCProfiler-SDK and ROCProfV3
 - Python bindings for rocprofiler-sdk-roctx
 - SQLite3 output support for rocprofv3 (`--output-format rocpd`)
 - Added `rocprofiler-sdk-rocpd` package
@@ -193,25 +193,31 @@ Full documentation for ROCprofiler-SDK is available at [rocm.docs.amd.com/projec
 - Added `rocpd` output format documentation
   - Requires the ROCprof Trace Decoder plugin installed (see above)
 - Added perfetto support for scratch memory.
+- Added documentation for rocprofv3 advanced options
 
 ### Changed
 
 - SDK no longer creates a background thread when every tool returns a nullptr from `rocprofiler_configure`.
 - Updated disassembly.hpp's vaddr-to-file-offset mapping to use the dedicated comgr API.
+- rocprofiler_uuid_t ABI is changed to hold 128 bit value.
 - rocprofv3 shorthand argument for `--collection-period` is now `-P` (upper-case) as `-p` (lower-case) is reserved for later use
 - default output format for rocprofv3 is now `rocpd` (SQLite3 database)
 - rocprofv3 avail tool renamed from rocprofv3_avail to rocprofv3-avail tool
 - rocprofv3 avail tool has support for command line arguments.
 - rocprofv3 tool now allows for Thread Trace + PC Sampling on the same agent
+- fixed inconsistency for what is a "null" handle in `rocprofiler_*_id_t` structs.
+  - correct answer is `.handle = 0` but some definitions used `UINT64_MAX`
 
 ### Resolved issues
 
 - Fixed missing callbacks around internal thread creation within counter collection service
 - Fixed potential data race in rocprofiler-sdk double buffering scheme
 - Usage of std::regex in core rocprofiler-sdk library which causes segfaults/exceptions when used under dual ABI
 - Fixed perfetto counter collection by introducing per dispatch accumulation.
+- Code object disassembly was missing function inlining information
 - Fixed queue preemption error and HSA_STATUS_ERROR_INVALID_PACKET_FORMAT error for stochastic PC-sampling for MI300X, leading to more stable runs.
 - Fixed the system hang issue for host-trap PC-sampling on MI300X. 
+- Fixed rocpd counter collection issue when counter collection alone is enabled, rocpd_kernel_dispatch table gets populated by counters data instead of kernel_dispatch data.
 
 ### Removed
 

README.md

@@ -68,7 +68,7 @@ cmake                                         \
       -DCMAKE_PREFIX_PATH=/opt/rocm           \
        rocprofiler-sdk-source
 
-  cmake --build rocprofiler-sdk-build --target all --parallel $(nproc)
+cmake --build rocprofiler-sdk-build --target all --parallel $(nproc)
 ```
 
 To install ROCprofiler, run:

requirements.txt

@@ -5,10 +5,12 @@ cmake>=3.21.0
 cmake-format
 dataclasses
 flake8
+jinja2
 numpy
 otf2
 pandas
 perfetto
 pycobertura
 pytest
 pyyaml
+reportlab

samples/CMakeLists.txt

@@ -47,3 +47,4 @@ add_subdirectory(code_object_isa_decode)
 add_subdirectory(external_correlation_id_request)
 add_subdirectory(pc_sampling)
 add_subdirectory(openmp_target)
+add_subdirectory(thread_trace)

samples/counter_collection/buffered_client.cpp

@@ -112,7 +112,7 @@ fill_dimension_cache(rocprofiler_counter_id_t counter)
     (*dimension_cache())
         ->emplace(counter.handle,
                   std::vector<rocprofiler_counter_record_dimension_info_t>{
-                      info.dimensions, info.dimensions + info.dimensions_count});
+                      *info.dimensions, *info.dimensions + info.dimensions_count});
 }
 
 /**

samples/counter_collection/device_counting_sync_client.cpp

@@ -294,7 +294,7 @@ counter_sampler::get_counter_size(rocprofiler_counter_id_t counter)
     ROCPROFILER_CALL(rocprofiler_query_counter_info(
                          counter, ROCPROFILER_COUNTER_INFO_VERSION_1, static_cast<void*>(&info)),
                      "Could not query info for counter");
-    return info.instance_ids_count;
+    return info.dimensions_instances_count;
 }
 
 std::unordered_map<std::string, rocprofiler_counter_id_t>
@@ -339,7 +339,7 @@ counter_sampler::get_counter_dimensions(rocprofiler_counter_id_t counter)
                          counter, ROCPROFILER_COUNTER_INFO_VERSION_1, static_cast<void*>(&info)),
                      "Could not query info for counter");
     return std::vector<rocprofiler_counter_record_dimension_info_t>{
-        info.dimensions, info.dimensions + info.dimensions_count};
+        *info.dimensions, *info.dimensions + info.dimensions_count};
 }
 
 std::atomic<bool>&

samples/counter_collection/print_functional_counters_client.cpp

@@ -292,15 +292,15 @@ dispatch_callback(rocprofiler_dispatch_counting_service_data_t dispatch_data,
                 "Could not query counter_id");
             cap.expected_counter_names.emplace(found_counter.handle, std::string(info.name));
             cap.remaining.push_back(found_counter);
-            cap.expected.emplace(found_counter.handle, info.instance_ids_count);
+            cap.expected.emplace(found_counter.handle, info.dimensions_instances_count);
 
             auto& info_vector =
                 cap.expected_data_dims.emplace(found_counter.handle, validate_dim_presence{})
                     .first->second;
 
             for(uint64_t i = 0; i < info.dimensions_count; i++)
             {
-                info_vector.maybe_forward(info.dimensions[i]);
+                info_vector.maybe_forward(*info.dimensions[i]);
             }
         }
         if(cap.expected.empty())