Collecting and Flushing Profile Data
You can call APIs to enable automatic collection of raw profile data. After the raw profile data is successfully collected, you can copy it to a development environment where the tool is installed to parse the data, and view the visualized parsing results.
API Overview
API |
Description |
|---|---|
aclprofCreateConfig |
Creates a profiling configuration. This API is used together with aclprofDestroyConfig. |
aclprofInit |
Initializes profiling and sets the path for saving profile data files. This API is used together with aclprofFinalize. |
aclprofSetConfig |
Functions as an extended API of aclprofCreateConfig and is used to configure collection parameters. |
aclprofStart |
Starts profiling. This API is used together with aclprofStop. |
aclprofStop |
Stops profiling. This API is used together with aclprofStart. |
aclprofFinalize |
Finalizes profiling. This API is used together with aclprofInit. |
aclprofDestroyConfig |
Destroys data of the aclprofConfig type created by the aclprofCreateConfig API call. This API is used together with aclprofCreateConfig. |
API Call Examples
The API call example is as follows:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | // 1. Call aclInit for initialization. // 2. Allocate runtime resources, including setting the compute device and creating a context and a stream. // 3. Initialize profiling. // Set the data flush path. const char *aclProfPath = "./output"; aclprofInit(aclProfPath, strlen(aclProfPath)); // 4. Configure profiling. uint32_t deviceIdList[1] = {0}; // Set this parameter based on the device ID in the actual environment. // Create a configuration structure. aclprofConfig *config = aclprofCreateConfig(deviceIdList, 1, ACL_AICORE_ARITHMETIC_UTILIZATION, nullptr,ACL_PROF_ACL_API | ACL_PROF_TASK_TIME); const char *memFreq = "15"; ret = aclprofSetConfig(ACL_PROF_SYS_HARDWARE_MEM_FREQ, memFreq, strlen(memFreq)); aclprofStart(config); // 5. Load the model. After the model is successfully loaded, modelId that identifies the model is returned. // 6. Create data of type aclmdlDataset to describe the inputs and outputs of the model. // 7. Execute the model. ret = aclmdlExecute(modelId, input, output); // 8. Process the model inference result. // 9. Destroy allocations such as the model inputs and outputs, free memory, and unload the model. // 10. Stop profiling and destroy the configuration and related resources. aclprofStop(config); aclprofDestroyConfig(config); aclprofFinalize(); // 11. Destroy runtime allocations. // 12. Call aclFinalize for deinitialization. //...... |