Improve and rename gpuinfo, simplify cl_hot_functions sample

jfedorov · Apr 14, 2021 · df968c9 · df968c9
1 parent 0f88019
commit df968c9
Show file tree

Hide file tree

Showing 18 changed files with 488 additions and 631 deletions.
diff --git a/README.md b/README.md
@@ -48,7 +48,7 @@ You may obtain a copy of the License at https://opensource.org/licenses/MIT
  - [onetrace](tools/onetrace) - host and device tracing tool for OpenCL(TM) and Level Zero backends with support of DPC++ (both for CPU and GPU) and OpenMP* GPU offload;
  - [ze_tracer](tools/ze_tracer) - "Swiss army knife" for Level Zero API call tracing and profiling (former ze_intercept);
  - [cl_tracer](tools/cl_tracer) - "Swiss army knife" for OpenCL(TM) API call tracing and profiling;
- - [gpu_info](tools/gpu_info) - provides basic information about the GPUs installed in a system, and the list of HW metrics one can collect for it;
+ - [gpuinfo](tools/gpuinfo) - provides basic information about the GPUs installed in a system, and the list of HW metrics one can collect for it;
 
 ## Sample Tools & Utilities
 - tools for OpenCL(TM), DPC++ (with OpenCL(TM) backend) and OpenMP* GPU offload (with OpenCL(TM) backend):

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-0.20.1
+0.21.0
diff --git a/chapters/device_activity_tracing/OpenCL.md b/chapters/device_activity_tracing/OpenCL.md
@@ -20,7 +20,7 @@ Intel(R) Xeon(R) Processor / Intel(R) Core(TM) Processor (CPU) Runtimes use `Que
 - Windows
 
 **Supported HW**:
-- Intel(R) Processor Graphics GEN9+
+- Any
 
 **Needed Headers**:
 - OpenCL(TM) [headers](https://github.com/KhronosGroup/OpenCL-Headers)

diff --git a/chapters/metrics_collection/MetricsDiscoveryAPI.md b/chapters/metrics_collection/MetricsDiscoveryAPI.md
@@ -295,5 +295,7 @@ LD_LIBRARY_PATH=$LD_LIBRARY_PATH:<path_to_libmd.so> ./<application>
 [Compute Architecture Manuals](https://software.intel.com/en-us/articles/intel-graphics-developers-guides) to learn more on Intel(R) Processor Graphics architecture
 
 ## Samples
-- [GPU Info](../../tools/gpu_info)
-- [GPU Metrics for OpenCL(TM)](../../samples/cl_gpu_metrics)
+- [GPU Metrics for OpenCL(TM)](../../samples/cl_gpu_metrics)
+
+## Tools
+- [GPU Info](../../tools/gpuinfo)
diff --git a/chapters/runtime_api_tracing/OpenCL.md b/chapters/runtime_api_tracing/OpenCL.md
@@ -49,7 +49,7 @@ cl_int CL_API_CALL clGetTracingStateINTEL(
 - Windows
 
 **Supported HW**:
-- Any
+- Intel(R) Processor Graphics, Intel(R) Xeon(R) Processor and Intel(R) Core(TM) Processor
 
 **Needed Headers**:
 - OpenCL(TM) [headers](https://github.com/KhronosGroup/OpenCL-Headers)
@@ -115,4 +115,8 @@ void Callback(cl_function_id fid,
 - [OpenCL(TM) Hot Functions](../../samples/cl_hot_functions)
 - [OpenCL(TM) Hot Kernels](../../samples/cl_hot_kernels)
 - [OpenCL(TM) Debug Info](../../samples/cl_debug_info)
-- [OpenCL(TM) GPU Metrics](../../samples/cl_gpu_metrics)
+- [OpenCL(TM) GPU Metrics](../../samples/cl_gpu_metrics)
+
+## Tools
+- [OpenCL(TM) Tracer](../../tools/cl_tracer)
+- [Tracing and Profiling Tool for Data Parallel C++ (DPC++)](../../tools/onetrace)
diff --git a/samples/cl_hot_functions/README.md b/samples/cl_hot_functions/README.md
@@ -1,43 +1,30 @@
 # OpenCL(TM) Hot Functions
 ## Overview
-This sample is a simple LD_PRELOAD based tool that allows to collect all called OpenCL(TM) API functions within an application along with their total execution time and call count.
+This is a simple LD_PRELOAD based tool that allows to collect all called OpenCL(TM) API functions within an application along with their total execution time and call count for GPU device.
 
 As a result, table like the following will be printed.
 ```
-=== API Timing Results: ===
-
-Total Execution Time (ns): 363687486
-Total API Time for CPU backend (ns): 524
-Total API Time for GPU backend (ns): 355355363
-
-== CPU Backend: ==
-
-      Function,       Calls,           Time (ns),  Time (%),        Average (ns),            Min (ns),            Max (ns)
-clGetDeviceIDs,           1,                 524,    100.00,                 524,                 524,                 524
-
-== GPU Backend: ==
-
-                          Function,       Calls,           Time (ns),  Time (%),        Average (ns),            Min (ns),            Max (ns)
-                    clBuildProgram,           1,           173888026,     48.93,           173888026,           173888026,           173888026
-                          clFinish,           4,           172908147,     48.66,            43227036,            42711785,            44318785
-              clEnqueueWriteBuffer,           8,             4636256,      1.30,              579532,              207825,             1864890
-               clEnqueueReadBuffer,           4,             2051244,      0.58,              512811,              498662,              542971
-            clEnqueueNDRangeKernel,           4,             1623139,      0.46,              405784,              236120,              609050
-                clReleaseMemObject,          12,               95182,      0.03,                7931,                3525,               16436
-                    clCreateBuffer,          12,               81056,      0.02,                6754,                2511,               16990
-                    clSetKernelArg,          16,               24515,      0.01,                1532,                 141,                7038
-           clGetEventProfilingInfo,           8,               13139,      0.00,                1642,                 103,                3288
-                   clCreateContext,           1,               12680,      0.00,               12680,               12680,               12680
-                  clReleaseProgram,           1,                9503,      0.00,                9503,                9503,                9503
-         clCreateProgramWithSource,           1,                3880,      0.00,                3880,                3880,                3880
-                    clCreateKernel,           1,                2941,      0.00,                2941,                2941,                2941
-                   clReleaseKernel,           1,                1679,      0.00,                1679,                1679,                1679
-                   clGetKernelInfo,           4,                1617,      0.00,                 404,                 190,                 552
-clCreateCommandQueueWithProperties,           1,                1388,      0.00,                1388,                1388,                1388
-                    clGetDeviceIDs,           2,                 311,      0.00,                 155,                 138,                 173
-             clReleaseCommandQueue,           1,                 270,      0.00,                 270,                 270,                 270
-                   clGetDeviceInfo,           2,                 227,      0.00,                 113,                 103,                 124
-                  clReleaseContext,           1,                 163,      0.00,                 163,                 163,                 163
+                          Function,       Calls,           Time (ns),        Average (ns)
+                    clBuildProgram,           1,           183549198,           183549198
+                    clCreateBuffer,          12,              108285,                9023
+clCreateCommandQueueWithProperties,           1,                1265,                1265
+                   clCreateContext,           1,                9322,                9322
+                    clCreateKernel,           1,                3428,                3428
+         clCreateProgramWithSource,           1,                3219,                3219
+            clEnqueueNDRangeKernel,           4,             2237845,              559461
+               clEnqueueReadBuffer,           4,             2358133,              589533
+              clEnqueueWriteBuffer,           8,             5719781,              714972
+                          clFinish,           4,           174064236,            43516059
+                    clGetDeviceIDs,           2,                 362,                 181
+                   clGetDeviceInfo,           2,                 354,                 177
+           clGetEventProfilingInfo,           8,               14198,                1774
+                   clGetKernelInfo,           4,                2411,                 602
+             clReleaseCommandQueue,           1,                1046,                1046
+                  clReleaseContext,           1,                 173,                 173
+                   clReleaseKernel,           1,                2741,                2741
+                clReleaseMemObject,          12,              110922,                9243
+                  clReleaseProgram,           1,               11561,               11561
+                    clSetKernelArg,          16,               75282,                4705
 ```
 ## Supported OS
 - Linux
@@ -48,8 +35,7 @@ clCreateCommandQueueWithProperties,           1,                1388,      0.00,
 - [Git](https://git-scm.com/) (version 1.8 and above)
 - [Python](https://www.python.org/) (version 2.7 and above)
 - [OpenCL(TM) ICD Loader](https://github.com/KhronosGroup/OpenCL-ICD-Loader)
-- [Intel(R) Graphics Compute Runtime for oneAPI Level Zero and OpenCL(TM) Driver](https://github.com/intel/compute-runtime) to run on GPU
-- [Intel(R) Xeon(R) Processor / Intel(R) Core(TM) Processor (CPU) Runtimes](https://software.intel.com/en-us/articles/opencl-drivers#cpu-section) to run on CPU
+- [Intel(R) Graphics Compute Runtime for oneAPI Level Zero and OpenCL(TM) Driver](https://github.com/intel/compute-runtime)
 
 ## Build and Run
 ### Linux
@@ -65,10 +51,9 @@ Use this command line to run the tool:
 ```sh
 ./cl_hot_functions <target_application>
 ```
-One may use [cl_gemm](../cl_gemm) or [dpc_gemm](../dpc_gemm) as target application:
+One may use [cl_gemm](../cl_gemm) as target application:
 ```sh
 ./cl_hot_functions ../../cl_gemm/build/cl_gemm
-./cl_hot_functions ../../dpc_gemm/build/dpc_gemm cpu
 ```
 ### Windows
 Use Microsoft* Visual Studio x64 command prompt to run the following commands and build the sample:
@@ -83,8 +68,7 @@ Use this command line to run the tool:
 ```sh
 cl_hot_functions.exe <target_application>
 ```
-One may use [cl_gemm](../cl_gemm) or [dpc_gemm](../dpc_gemm) as target application:
+One may use [cl_gemm](../cl_gemm) as target application:
 ```sh
 cl_hot_functions.exe ..\..\cl_gemm\build\cl_gemm.exe
-cl_hot_functions.exe ..\..\dpc_gemm\build\dpc_gemm.exe cpu
 ```
diff --git a/samples/cl_hot_functions/cl_api_collector.h b/samples/cl_hot_functions/cl_api_collector.h