Welton, Benjamin
aa88dd44c7
Adds iteration based multiplexing to counter collection. Counter groups can now be specified. These counter groups are collected on a device individually until a specified interval period is reached. When the interval is reached, the next counter group is set to be collected on subsequent kernel executions. Supplies two new argument types that can be included in YAML/JSON inputs: pmc_groups: an array of arrays containing the counter groups to run (i.e. [ ["SQ_WAVES", "GRBM_COUNT"], ["GRBM_GUI_ACTIVE"]) pmc_group_interval: the number of kernel invocations on a GPU of a group before rotating to the next group Note: originally there was a random_seed_generator proposed in the linked ticket, that was not implemented since there are very few instances where you would want the selection of the groups to be randomly generated (and if you do, you can randomly generate the pattern and place it as a large list of groups in pmc_group). All existing counter functionality should be preserved (selection of counters on specific devices only, profiling of only specific kernels, etc). --------- Co-authored-by: Benjamin Welton <bewelton@amd.com>
190 linhas
7.5 KiB
JSON
190 linhas
7.5 KiB
JSON
{
|
|
"$schema": "http://json-schema.org/draft-07/schema#",
|
|
"type": "object",
|
|
"description": "Schema for the rocprofv3 JSON input",
|
|
"properties": {
|
|
"jobs": {
|
|
"type": "array",
|
|
"description": "rocprofv3 input data per application run",
|
|
"items": {
|
|
"type" : "object",
|
|
"description": "data for rocprofv3",
|
|
"properties": {
|
|
|
|
"pmc": {
|
|
"type" : "array",
|
|
"description": "list of counters to collect"
|
|
},
|
|
|
|
"pmc_groups": {
|
|
"type" : "array",
|
|
"description": "An array containing lists of PMC counters to collect in a multiplexing fashion (e.x. [[counter1, counter2], [counter3, counter4]])"
|
|
},
|
|
"pmc_group_interval": {
|
|
"type" : "integer",
|
|
"description": "Number of kernel launches between selecting the next group of counters to collect"
|
|
},
|
|
|
|
"kernel_include_regex":{
|
|
"type": "string",
|
|
"description": "Include the kernels matching this filter"
|
|
},
|
|
|
|
"kernel_exclude_regex": {
|
|
"type": "string",
|
|
"description": "Exclude the kernels matching this filter"
|
|
},
|
|
|
|
"kernel_iteration_range": {
|
|
"type": "string",
|
|
"description": "Iteration range for each kernel that match the filter [start-stop]"
|
|
},
|
|
|
|
"hip_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HIP Traces (runtime + compiler)"
|
|
},
|
|
|
|
"hip_runtime_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HIP Runtime API Traces"
|
|
},
|
|
|
|
"hip_compiler_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HIP Compiler generated code Traces"
|
|
},
|
|
|
|
"marker_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting Marker (ROCTx) Traces"
|
|
},
|
|
|
|
"kernel_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting Kernel Dispatch Traces"
|
|
},
|
|
|
|
"memory_copy_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting Memory Copy Traces"
|
|
},
|
|
|
|
"memory_allocation_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting Memory Allocation Traces"
|
|
},
|
|
"rocdecode_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting rocDecode Traces"
|
|
},
|
|
"rocjpeg_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting rocJPEG Traces"
|
|
},
|
|
"scratch_memory_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting Scratch Memory operations Traces"
|
|
},
|
|
|
|
"stats": {
|
|
"type": "boolean",
|
|
"description": "For Collecting statistics of enabled tracing types"
|
|
},
|
|
|
|
"hsa_trace": {
|
|
"type": "boolean",
|
|
"description":"For Collecting HSA Traces (core + amd + image + finalizer)"
|
|
},
|
|
|
|
"hsa_core_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HSA API Traces (core API)"
|
|
},
|
|
|
|
"hsa_amd_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HSA API Traces (AMD-extension API)"
|
|
},
|
|
|
|
"hsa_finalize_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HSA API Traces (Finalizer-extension API)"
|
|
},
|
|
|
|
"hsa_image_trace": {
|
|
"type": "boolean",
|
|
"description": "For Collecting HSA API Traces (Image-extenson API)"
|
|
},
|
|
|
|
"runtime_trace" : {
|
|
"type": "boolean",
|
|
"description": "For collecting HIP (runtime), memory copy, memory allocation, marker, scratch memory, rocDecode, and Kernel dispatch traces."
|
|
},
|
|
|
|
"sys_trace" : {
|
|
"type": "boolean",
|
|
"description": "For Collecting HIP, HSA, Marker (ROCTx), Memory copy, Memory allocation, Scratch memory, rocDecode, rocJPEG, and Kernel dispatch traces"
|
|
},
|
|
|
|
"mangled_kernels": {
|
|
"type": "boolean",
|
|
"description": "Do not demangle the kernel names"
|
|
},
|
|
|
|
"truncate_kernels": {
|
|
"type": "boolean",
|
|
"description": "Truncate the demangled kernel names"
|
|
},
|
|
|
|
"output_file":{
|
|
"type": "string",
|
|
"description": "For the output file name"
|
|
},
|
|
|
|
"output_directory":{
|
|
"type": "string",
|
|
"description": "For adding output path where the output files will be saved"
|
|
},
|
|
|
|
"output_format":{
|
|
"type": "array",
|
|
"description": "For adding output format (supported formats: csv, json, pftrace)"
|
|
},
|
|
|
|
"list_metrics" : {
|
|
"type" : "boolean",
|
|
"description": "List the metrics"
|
|
},
|
|
|
|
"log_level":{
|
|
"type": "string",
|
|
"description": "fatal, error, warning, info, trace"
|
|
},
|
|
|
|
"preload":{
|
|
"type": "array",
|
|
"description": "Libraries to prepend to LD_PRELOAD (usually for sanitizers)"
|
|
},
|
|
"pc_sampling_unit": {
|
|
"type": "string",
|
|
"description": "pc sampling unit"
|
|
},
|
|
"pc_sampling_method": {
|
|
"type": "string",
|
|
"description": "pc sampling method"
|
|
},
|
|
"pc_sampling_interval": {
|
|
"type": "integer",
|
|
"description": "pc sampling interval"
|
|
},
|
|
"pc-sampling-beta-enabled": {
|
|
"type": "boolean",
|
|
"description": "enable pc sampling support; beta version"
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|