AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Removing ATT buffer size limitation (#534)

Browse Source 
* Removing SQTT buffer size limitation

* Update source/lib/rocprofiler-sdk/thread_trace/core.cpp

* Added testing for buffer size. Formatting.

* Add test as unstable

* Increase default buffer size

* Apply suggestions from code review

Co-authored-by: Indic, Vladimir <Vladimir.Indic@amd.com>

* Fix typo from code review

* Update tests/thread-trace/agent.cpp

---------

Co-authored-by: Giovanni <gbaraldi@amd.com>
Co-authored-by: Indic, Vladimir <Vladimir.Indic@amd.com>
This commit is contained in:
Baraldi, Giovanni
2025-07-29 22:47:40 +02:00
committed by GitHub
parent 2d8936362e
commit 1ba08cd4df
8 changed files with 53 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files
source/lib/rocprofiler-sdk/aql/packet_construct.cpp
+6 -2
View File
@@ -145,7 +145,8 @@ ThreadTraceAQLPacketFactory::ThreadTraceAQLPacketFactory(const hsa::AgentCache&
uint32_t cu = static_cast<uint32_t>(params.target_cu);
uint32_t shader_engine_mask = static_cast<uint32_t>(params.shader_engine_mask);
uint32_t simd = static_cast<uint32_t>(params.simd_select);
uint32_t buffer_size = static_cast<uint32_t>(params.buffer_size);
uint32_t buffer_size_lo = static_cast<uint32_t>(params.buffer_size);
uint32_t buffer_size_hi = static_cast<uint32_t>(params.buffer_size >> 32);
uint32_t perf_ctrl = static_cast<uint32_t>(params.perfcounter_ctrl);
aql_params.clear();
@@ -153,7 +154,10 @@ ThreadTraceAQLPacketFactory::ThreadTraceAQLPacketFactory(const hsa::AgentCache&
aql_params.push_back({HSA_VEN_AMD_AQLPROFILE_PARAMETER_NAME_COMPUTE_UNIT_TARGET, {cu}});
aql_params.push_back({HSA_VEN_AMD_AQLPROFILE_PARAMETER_NAME_SE_MASK, {shader_engine_mask}});
aql_params.push_back({HSA_VEN_AMD_AQLPROFILE_PARAMETER_NAME_SIMD_SELECTION, {simd}});
aql_params.push_back({HSA_VEN_AMD_AQLPROFILE_PARAMETER_NAME_ATT_BUFFER_SIZE, {buffer_size}});
aql_params.push_back({HSA_VEN_AMD_AQLPROFILE_PARAMETER_NAME_ATT_BUFFER_SIZE, {buffer_size_lo}});
if(buffer_size_hi != 0) aql_params.push_back({static_cast<hsa_ven_amd_aqlprofile_parameter_name_t>(
AQLPROFILE_ATT_PARAMETER_NAME_BUFFER_SIZE_HIGH), {buffer_size_hi}});
if(perf_ctrl != 0 && !params.perfcounters.empty())
{