// // Copyright (c) 2017 The Khronos Group Inc. // // Licensed under the Apache License, Version 2.0 (the "License"); // you may not use this file except in compliance with the License. // You may obtain a copy of the License at // // http://www.apache.org/licenses/LICENSE-2.0 // // Unless required by applicable law or agreed to in writing, software // distributed under the License is distributed on an "AS IS" BASIS, // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. // See the License for the specific language governing permissions and // limitations under the License. // #include "testBase.h" // This test is designed to stress changing kernel arguments between execute calls (that are asynchronous and thus // potentially overlapping) to make sure each kernel gets the right arguments // Note: put a delay loop in the kernel to make sure we have time to queue the next kernel before this one finishes const char *inspect_image_kernel_source[] = { "__kernel void sample_test(read_only image2d_t src, __global int *outDimensions )\n" "{\n" " int tid = get_global_id(0), i;\n" " for( i = 0; i < 100000; i++ ); \n" " outDimensions[tid * 2] = get_image_width(src) * tid;\n" " outDimensions[tid * 2 + 1] = get_image_height(src) * tid;\n" "\n" "}\n" }; #define NUM_TRIES 100 #define NUM_THREADS 2048 int test_kernel_arg_changes(cl_device_id device, cl_context context, cl_command_queue queue, int num_elements) { clProgramWrapper program; clKernelWrapper kernel; int error, i; clMemWrapper images[ NUM_TRIES ]; size_t sizes[ NUM_TRIES ][ 2 ]; clMemWrapper results[ NUM_TRIES ]; cl_image_format imageFormat; size_t maxWidth, maxHeight; size_t threads[1], localThreads[1]; cl_int resultArray[ NUM_THREADS * 2 ]; char errStr[ 128 ]; RandomSeed seed( gRandomSeed ); PASSIVE_REQUIRE_IMAGE_SUPPORT( device ) // Just get any ol format to test with error = get_8_bit_image_format( context, CL_MEM_OBJECT_IMAGE2D, CL_MEM_READ_WRITE, 0, &imageFormat ); test_error( error, "Unable to obtain suitable image format to test with!" ); // Create our testing kernel error = create_single_kernel_helper( context, &program, &kernel, 1, inspect_image_kernel_source, "sample_test" ); test_error( error, "Unable to create testing kernel" ); // Get max dimensions for each of our images error = clGetDeviceInfo( device, CL_DEVICE_IMAGE2D_MAX_WIDTH, sizeof( maxWidth ), &maxWidth, NULL ); error |= clGetDeviceInfo( device, CL_DEVICE_IMAGE2D_MAX_HEIGHT, sizeof( maxHeight ), &maxHeight, NULL ); test_error( error, "Unable to get max image dimensions for device" ); // Get the number of threads we'll be able to run threads[0] = NUM_THREADS; error = get_max_common_work_group_size( context, kernel, threads[0], &localThreads[0] ); test_error( error, "Unable to get work group size for kernel" ); // Create a variety of images and output arrays for( i = 0; i < NUM_TRIES; i++ ) { sizes[ i ][ 0 ] = genrand_int32(seed) % (maxWidth/32) + 1; sizes[ i ][ 1 ] = genrand_int32(seed) % (maxHeight/32) + 1; images[i] = create_image_2d(context, CL_MEM_READ_ONLY, &imageFormat, sizes[i][0], sizes[i][1], 0, NULL, &error); if( images[i] == NULL ) { log_error("Failed to create image %d of size %d x %d (%s).\n", i, (int)sizes[i][0], (int)sizes[i][1], IGetErrorString( error )); return -1; } results[i] = clCreateBuffer(context, CL_MEM_READ_WRITE, sizeof(cl_int) * threads[0] * 2, NULL, &error); if( results[i] == NULL) { log_error("Failed to create array %d of size %d.\n", i, (int)threads[0]*2); return -1; } } // Start setting arguments and executing kernels for( i = 0; i < NUM_TRIES; i++ ) { // Set the arguments for this try error = clSetKernelArg( kernel, 0, sizeof( cl_mem ), &images[ i ] ); sprintf( errStr, "Unable to set argument 0 for kernel try %d", i ); test_error( error, errStr ); error = clSetKernelArg( kernel, 1, sizeof( cl_mem ), &results[ i ] ); sprintf( errStr, "Unable to set argument 1 for kernel try %d", i ); test_error( error, errStr ); // Queue up execution error = clEnqueueNDRangeKernel( queue, kernel, 1, NULL, threads, localThreads, 0, NULL, NULL ); sprintf( errStr, "Unable to execute kernel try %d", i ); test_error( error, errStr ); } // Read the results back out, one at a time, and verify for( i = 0; i < NUM_TRIES; i++ ) { error = clEnqueueReadBuffer( queue, results[ i ], CL_TRUE, 0, sizeof( cl_int ) * threads[0] * 2, resultArray, 0, NULL, NULL ); sprintf( errStr, "Unable to read results for kernel try %d", i ); test_error( error, errStr ); // Verify. Each entry should be n * the (width/height) of image i for( int j = 0; j < NUM_THREADS; j++ ) { if( resultArray[ j * 2 + 0 ] != (int)sizes[ i ][ 0 ] * j ) { log_error( "ERROR: Verficiation for kernel try %d, sample %d FAILED, expected a width of %d, got %d\n", i, j, (int)sizes[ i ][ 0 ] * j, resultArray[ j * 2 + 0 ] ); return -1; } if( resultArray[ j * 2 + 1 ] != (int)sizes[ i ][ 1 ] * j ) { log_error( "ERROR: Verficiation for kernel try %d, sample %d FAILED, expected a height of %d, got %d\n", i, j, (int)sizes[ i ][ 1 ] * j, resultArray[ j * 2 + 1 ] ); return -1; } } } // If we got here, everything verified successfully return 0; }