一.背景
- 希望对比RK3588上CPU和Mali-GPU的性能差异
- Mali-GPU算力测试采用clpeak
- CPU-FP32的性能测试采用Openblas(开启了NEON优化)
- 分别用CPU和opencl测试opencv resize在不同算法下的性能:从32x32放大到8192x8192再缩放回32x32,循环100次
二.小结
- GPU型号: Mali-LODX r0p0 Mali-G610 4 cores r0p0 0xA867
- GPU FP32(clpeak): 441.95 GFLOPS
- CPU FP32(openblas+neon): 53.68 GFLOPS
- 插值方法：INTER_NEAREST CPU耗时(秒):3.01526 GPU耗时(秒):0.0672681
- 插值方法：INTER_LINEAR CPU耗时(秒):5.3227 GPU耗时(秒):0.0189366
- 插值方法：INTER_CUBIC CPU耗时(秒):8.22734 GPU耗时(秒):11.6337
- 插值方法：INTER_AREA CPU耗时(秒):20.4999 GPU耗时(秒):27.3197
- 插值方法：INTER_LANCZOS4 CPU耗时(秒):29.3602 GPU耗时(秒):43.9484
三.相关链接
- opencv编译
四.操作步骤
1.环境搭建
A.安装依赖
mv /lib/aarch64-linux-gnu/libOpenCL.so.1 /lib/aarch64-linux-gnu/libOpenCL.so.1.bk
ln -s /usr/lib/aarch64-linux-gnu/libmali.so /lib/aarch64-linux-gnu/libOpenCL.so.1sudo apt install opencl-headers
sudo apt install ocl-icd-libopencl1
sudo apt install ocl-icd-opencl-dev
sudo apt install clinfo
GPU_35">B.设置GPU为高性能模式
echo performance> /sys/class/devfreq/fb000000.gpu/governor
echo performance> /sys/class/devfreq/fdab0000.npu/governor
GPU_41">C.获取GPU信息
cat /sys/class/misc/mali0/device/gpuinfo
clinfo
输出
Mali-G610 4 cores r0p0 0xA867Number of platforms 1Platform Name ARM PlatformPlatform Vendor ARMPlatform Version OpenCL 2.1 v1.g6p0-01eac0.ba52c908d926792b8f5fe28f383a2b03Platform Profile FULL_PROFILEPlatform Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_3d_image_writes cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_fp16 cl_khr_icd cl_khr_egl_image cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_subgroups cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_il_program cl_khr_priority_hints cl_khr_create_command_queue cl_khr_spirv_no_integer_wrap_decoration cl_khr_extended_versioning cl_khr_device_uuid cl_arm_core_id cl_arm_printf cl_arm_non_uniform_work_group_size cl_arm_import_memory cl_arm_import_memory_dma_buf cl_arm_import_memory_host cl_arm_integer_dot_product_int8 cl_arm_integer_dot_product_accumulate_int8 cl_arm_integer_dot_product_accumulate_saturate_int8 cl_arm_scheduling_controls cl_arm_controlled_kernel_termination cl_ext_cxx_for_openclPlatform Host timer resolution 1nsPlatform Extensions function suffix ARMPlatform Name ARM Platform
Number of devices 1
arm_release_ver of this libmali is 'g6p0-01eac0', rk_so_ver is '6'.Device Name Mali-LODX r0p0Device Vendor ARMDevice Vendor ID 0xa8670000Device Version OpenCL 2.1 v1.g6p0-01eac0.ba52c908d926792b8f5fe28f383a2b03Driver Version 2.1Device OpenCL C Version OpenCL C 2.0 v1.g6p0-01eac0.ba52c908d926792b8f5fe28f383a2b03Device Type GPUDevice Profile FULL_PROFILEDevice Available YesCompiler Available YesLinker Available YesMax compute units 4Max clock frequency 1000MHzDevice Partition (core)Max number of sub-devices 0Supported partition types NoneSupported affinity domains (n/a)Max work item dimensions 3Max work item sizes 1024x1024x1024Max work group size 1024Preferred work group size multiple 16Max sub-groups per work group 64Preferred / native vector sizeschar 16 / 4short 8 / 2int 4 / 1long 2 / 1half 8 / 2 (cl_khr_fp16)float 4 / 1double 0 / 0 (n/a)Half-precision Floating-point support (cl_khr_fp16)Denormals YesInfinity and NANs YesRound to nearest YesRound to zero YesRound to infinity YesIEEE754-2008 fused multiply-add YesSupport is emulated in software NoSingle-precision Floating-point support (core)Denormals YesInfinity and NANs YesRound to nearest YesRound to zero YesRound to infinity YesIEEE754-2008 fused multiply-add YesSupport is emulated in software NoCorrectly-rounded divide and sqrt operations NoDouble-precision Floating-point support (n/a)Address bits 64, Little-EndianGlobal memory size 16643870720 (15.5GiB)Error Correction support NoMax memory allocation 16643870720 (15.5GiB)Unified memory for Host and Device YesShared Virtual Memory (SVM) capabilities (core)Coarse-grained buffer sharing YesFine-grained buffer sharing NoFine-grained system sharing NoAtomics NoMinimum alignment for any data type 128 bytesAlignment of base address 1024 bits (128 bytes)Preferred alignment for atomicsSVM 0 bytesGlobal 0 bytesLocal 0 bytesMax size for global variable 65536 (64KiB)Preferred total size of global vars 0Global Memory cache type Read/WriteGlobal Memory cache size 1048576 (1024KiB)Global Memory cache line size 64 bytesImage support YesMax number of samplers per kernel 16Max size for 1D images from buffer 65536 pixelsMax 1D or 2D image array size 2048 imagesBase address alignment for 2D image buffers 32 bytesPitch alignment for 2D image buffers 64 pixelsMax 2D image size 65536x65536 pixelsMax 3D image size 65536x65536x65536 pixelsMax number of read image args 128Max number of write image args 64Max number of read/write image args 64Max number of pipe args 16Max active pipe reservations 1Max pipe packet size 1024Local memory type GlobalLocal memory size 32768 (32KiB)Max number of constant args 128Max constant buffer size 16643870720 (15.5GiB)Max size of kernel argument 1024Queue properties (on host)Out-of-order execution YesProfiling YesQueue properties (on device)Out-of-order execution YesProfiling YesPreferred size 2097152 (2MiB)Max size 16777216 (16MiB)Max queues on device 1Max events on device 1024Prefer user sync for interop NoProfiling timer resolution 1000nsExecution capabilitiesRun OpenCL kernels YesRun native kernels NoSub-group independent forward progress YesIL version SPIR-V_1.0SPIR versions <printDeviceInfo:161: get CL_DEVICE_SPIR_VERSIONS size : error -30>printf() buffer size 1048576 (1024KiB)Built-in kernels (n/a)Device Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_3d_image_writes cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_fp16 cl_khr_icd cl_khr_egl_image cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_subgroups cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_il_program cl_khr_priority_hints cl_khr_create_command_queue cl_khr_spirv_no_integer_wrap_decoration cl_khr_extended_versioning cl_khr_device_uuid cl_arm_core_id cl_arm_printf cl_arm_non_uniform_work_group_size cl_arm_import_memory cl_arm_import_memory_dma_buf cl_arm_import_memory_host cl_arm_integer_dot_product_int8 cl_arm_integer_dot_product_accumulate_int8 cl_arm_integer_dot_product_accumulate_saturate_int8 cl_arm_scheduling_controls cl_arm_controlled_kernel_termination cl_ext_cxx_for_openclNULL platform behaviorclGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) ARM PlatformclGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [ARM]clCreateContext(NULL, ...) [default] Success [ARM]clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) Success (1)Platform Name ARM PlatformDevice Name Mali-LODX r0p0clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platformclCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)Platform Name ARM PlatformDevice Name Mali-LODX r0p0clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platformclCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) No devices found in platformclCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) Success (1)Platform Name ARM PlatformDevice Name Mali-LODX r0p0
D.获取CPU信息
lscpu
输出
Architecture: aarch64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 8
On-line CPU(s) list: 0-7
Thread(s) per core: 1
Core(s) per socket: 2
Socket(s): 3
Vendor ID: ARM
Model: 0
Model name: Cortex-A55
Stepping: r2p0
CPU max MHz: 2208.0000
CPU min MHz: 408.0000
BogoMIPS: 48.00
L1d cache: 256 KiB
L1i cache: 256 KiB
L2 cache: 1 MiB
L3 cache: 3 MiB
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp
GPU_229">2.调用OpenCL SDK获取GPU信息
cat > cl_query.c <<-'EOF'
#include <stdio.h>
#include <stdlib.h>
#include <CL/cl.h>int main() {cl_platform_id *platforms = NULL;cl_uint num_platforms = 0;// 获取可用的平台数量cl_int clStatus = clGetPlatformIDs(0, NULL, &num_platforms);platforms = (cl_platform_id*) malloc(sizeof(cl_platform_id) * num_platforms);// 获取所有平台IDclStatus = clGetPlatformIDs(num_platforms, platforms, NULL);printf("OpenCL平台数量: %d\n", num_platforms);// 遍历每个平台for (cl_uint i = 0; i < num_platforms; ++i) {char buffer[10240];printf("\n平台 %d:\n", i+1);// 获取平台名称clGetPlatformInfo(platforms[i], CL_PLATFORM_NAME, sizeof(buffer), buffer, NULL);printf(" 名称: %s\n", buffer);// 获取平台供应商clGetPlatformInfo(platforms[i], CL_PLATFORM_VENDOR, sizeof(buffer), buffer, NULL);printf(" 供应商: %s\n", buffer);// 获取平台版本clGetPlatformInfo(platforms[i], CL_PLATFORM_VERSION, sizeof(buffer), buffer, NULL);printf(" 版本: %s\n", buffer);// 获取设备数量cl_uint num_devices = 0;clGetDeviceIDs(platforms[i], CL_DEVICE_TYPE_ALL, 0, NULL, &num_devices);cl_device_id *devices = (cl_device_id*) malloc(sizeof(cl_device_id) * num_devices);clGetDeviceIDs(platforms[i], CL_DEVICE_TYPE_ALL, num_devices, devices, NULL);// 遍历每个设备for (cl_uint j = 0; j < num_devices; ++j) {printf(" 设备 %d:\n", j+1);// 获取设备名称clGetDeviceInfo(devices[j], CL_DEVICE_NAME, sizeof(buffer), buffer, NULL);printf(" 名称: %s\n", buffer);// 获取设备类型cl_device_type device_type;clGetDeviceInfo(devices[j], CL_DEVICE_TYPE, sizeof(device_type), &device_type, NULL);if (device_type & CL_DEVICE_TYPE_CPU)printf(" 类型: CPU\n");if (device_type & CL_DEVICE_TYPE_GPU)printf(" 类型: GPU\n");if (device_type & CL_DEVICE_TYPE_ACCELERATOR)printf(" 类型: 加速器\n");// 获取计算单元数量cl_uint compute_units;clGetDeviceInfo(devices[j], CL_DEVICE_MAX_COMPUTE_UNITS, sizeof(compute_units), &compute_units, NULL);printf(" 计算单元数: %d\n", compute_units);// 获取全局内存大小cl_ulong global_mem;clGetDeviceInfo(devices[j], CL_DEVICE_GLOBAL_MEM_SIZE, sizeof(global_mem), &global_mem, NULL);printf(" 全局内存大小: %llu MB\n", (unsigned long long)(global_mem / (1024 * 1024)));}free(devices);}free(platforms);return 0;
}
EOFgcc -o cl_query cl_query.c -lOpenCL
./cl_query
输出
OpenCL平台数量: 1平台 1:名称: ARM Platform供应商: ARM版本: OpenCL 2.1 v1.g6p0-01eac0.ba52c908d926792b8f5fe28f383a2b03设备 1:
arm_release_ver of this libmali is 'g6p0-01eac0', rk_so_ver is '6'.名称: Mali-LODX r0p0类型: GPU计算单元数: 4全局内存大小: 15872 MB
3.使用OpenCL API计算矩阵乘
cat > matmul.c <<-'EOF'
#include <stdio.h>
#include <stdlib.h>
#include <CL/cl.h>
#include <time.h>
#include <sys/time.h>#define MATRIX_SIZE 8192
#define TILE_SIZE 32// 获取当前时间（秒），用于计算耗时
double get_current_time() {struct timeval tp;gettimeofday(&tp, NULL);return (double)(tp.tv_sec) + (double)(tp.tv_usec) / 1e6;
}#define xstr(s) str(s)
#define str(s) #sconst char *kernelSource = " \n" \
"__kernel void mat_mul_optimized(const int N, \n" \
" __global float* A, \n" \
" __global float* B, \n" \
" __global float* C) { \n" \
" const int TILE_SIZE = " xstr(TILE_SIZE) "; \n" \
" __local float Asub[TILE_SIZE][TILE_SIZE]; \n" \
" __local float Bsub[TILE_SIZE][TILE_SIZE]; \n" \
" int global_row = get_global_id(1); \n" \
" int global_col = get_global_id(0); \n" \
" int local_row = get_local_id(1); \n" \
" int local_col = get_local_id(0); \n" \
" float sum = 0.0f; \n" \
" int numTiles = (N + TILE_SIZE - 1) / TILE_SIZE; \n" \
" for (int t = 0; t < numTiles; ++t) { \n" \
" int tiled_row = global_row; \n" \
" int tiled_col = t * TILE_SIZE + local_col; \n" \
" if (tiled_row < N && tiled_col < N) \n" \
" Asub[local_row][local_col] = A[tiled_row * N + tiled_col];\n" \
" else \n" \
" Asub[local_row][local_col] = 0.0f; \n" \
" tiled_row = t * TILE_SIZE + local_row; \n" \
" tiled_col = global_col; \n" \
" if (tiled_row < N && tiled_col < N) \n" \
" Bsub[local_row][local_col] = B[tiled_row * N + tiled_col];\n" \
" else \n" \
" Bsub[local_row][local_col] = 0.0f; \n" \
" barrier(CLK_LOCAL_MEM_FENCE); \n" \
" for (int k = 0; k < TILE_SIZE; ++k) { \n" \
" sum += Asub[local_row][k] * Bsub[k][local_col]; \n" \
" } \n" \
" barrier(CLK_LOCAL_MEM_FENCE); \n" \
" } \n" \
" if (global_row < N && global_col < N) \n" \
" C[global_row * N + global_col] = sum; \n" \
"} \n";int main() {int N = MATRIX_SIZE;size_t bytes = N * N * sizeof(float);// 分配主机内存float *h_A = (float*)malloc(bytes);float *h_B = (float*)malloc(bytes);float *h_C = (float*)malloc(bytes);// 初始化矩阵for(int i = 0; i < N*N; i++) {h_A[i] = 1.0f;h_B[i] = 1.0f;}// 获取平台和设备信息cl_platform_id platformId = NULL;cl_device_id deviceID = NULL;cl_uint retNumDevices;cl_uint retNumPlatforms;cl_int ret = clGetPlatformIDs(1, &platformId, &retNumPlatforms);ret = clGetDeviceIDs(platformId, CL_DEVICE_TYPE_DEFAULT, 1, &deviceID, &retNumDevices);// 创建 OpenCL 上下文cl_context context = clCreateContext(NULL, 1, &deviceID, NULL, NULL, &ret);// 创建命令队列cl_command_queue commandQueue = clCreateCommandQueue(context, deviceID, 0, &ret);// 创建内存缓冲区cl_mem d_A = clCreateBuffer(context, CL_MEM_READ_ONLY, bytes, NULL, &ret);cl_mem d_B = clCreateBuffer(context, CL_MEM_READ_ONLY, bytes, NULL, &ret);cl_mem d_C = clCreateBuffer(context, CL_MEM_WRITE_ONLY, bytes, NULL, &ret);// 将数据写入缓冲区ret = clEnqueueWriteBuffer(commandQueue, d_A, CL_TRUE, 0, bytes, h_A, 0, NULL, NULL);ret = clEnqueueWriteBuffer(commandQueue, d_B, CL_TRUE, 0, bytes, h_B, 0, NULL, NULL);// 记录编译开始时间double compile_start = get_current_time();// 创建程序对象cl_program program = clCreateProgramWithSource(context, 1, (const char**)&kernelSource, NULL, &ret);// 编译内核程序ret = clBuildProgram(program, 1, &deviceID, NULL, NULL, NULL);// 检查编译错误if (ret != CL_SUCCESS) {size_t log_size;clGetProgramBuildInfo(program, deviceID, CL_PROGRAM_BUILD_LOG, 0, NULL, &log_size);char *log = (char *)malloc(log_size);clGetProgramBuildInfo(program, deviceID, CL_PROGRAM_BUILD_LOG, log_size, log, NULL);printf("CL Compilation failed:\n%s\n", log);free(log);return 1;}// 记录编译结束时间double compile_end = get_current_time();double compile_time = compile_end - compile_start;// 创建 OpenCL 内核cl_kernel kernel = clCreateKernel(program, "mat_mul_optimized", &ret);// 设置内核参数ret = clSetKernelArg(kernel, 0, sizeof(int), (void*)&N);ret = clSetKernelArg(kernel, 1, sizeof(cl_mem), (void*)&d_A);ret = clSetKernelArg(kernel, 2, sizeof(cl_mem), (void*)&d_B);ret = clSetKernelArg(kernel, 3, sizeof(cl_mem), (void*)&d_C);// 定义全局和本地工作区大小size_t local[2] = {TILE_SIZE, TILE_SIZE};size_t global[2] = {(size_t)((N + TILE_SIZE - 1) / TILE_SIZE) * TILE_SIZE,(size_t)((N + TILE_SIZE - 1) / TILE_SIZE) * TILE_SIZE};// 记录第一次内核执行开始时间double launch_start = get_current_time();// 执行内核ret = clEnqueueNDRangeKernel(commandQueue, kernel, 2, NULL, global, local, 0, NULL, NULL);printf("clEnqueueNDRangeKernel:%d\n",ret);// 等待命令队列执行完成clFinish(commandQueue);// 记录第一次内核执行结束时间double launch_end = get_current_time();double launch_time = launch_end - launch_start;// 读取结果ret = clEnqueueReadBuffer(commandQueue, d_C, CL_TRUE, 0, bytes, h_C, 0, NULL, NULL);// 计算 GFLOPSdouble total_ops = 2.0 * N * N * N;double gflops = (total_ops / 1e9) / launch_time;// 输出结果printf("编译时间: %f 秒\n", compile_time);printf("第一次内核执行时间: %f 秒\n", launch_time);printf("计算性能: %f GFLOPS\n", gflops);// 释放资源ret = clFlush(commandQueue);ret = clFinish(commandQueue);ret = clReleaseKernel(kernel);ret = clReleaseProgram(program);ret = clReleaseMemObject(d_A);ret = clReleaseMemObject(d_B);ret = clReleaseMemObject(d_C);ret = clReleaseCommandQueue(commandQueue);ret = clReleaseContext(context);free(h_A);free(h_B);free(h_C);return 0;
}EOF
gcc -o matmul matmul.c -lOpenCL
./matmul
输出
编译时间: 0.031085 秒
第一次内核执行时间: 62.258528 秒
计算性能: 17.660418 GFLOPS
GPU_521">4.使用clpeak测试GPU的性能
git clone https://gitcode.com/gh_mirrors/cl/clpeak.git
git submodule update --init --recursive --remote
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .
./clpeak
输出
Platform: ARM Platform
arm_release_ver of this libmali is 'g6p0-01eac0', rk_so_ver is '6'.Device: Mali-LODX r0p0Driver version : 2.1 (Linux ARM64)Compute units : 4Clock frequency : 1000 MHzGlobal memory bandwidth (GBPS)float : 25.71float2 : 24.45float4 : 23.70float8 : 12.05float16 : 12.01Single-precision compute (GFLOPS)float : 441.77float2 : 470.27float4 : 466.52float8 : 435.65float16 : 411.38Half-precision compute (GFLOPS)half : 441.96half2 : 878.25half4 : 911.51half8 : 886.19half16 : 846.44No double precision support! SkippedInteger compute (GIOPS)int : 124.96int2 : 125.71int4 : 125.16int8 : 123.82int16 : 124.24Integer compute Fast 24bit (GIOPS)int : 125.16int2 : 125.63int4 : 125.20int8 : 123.73int16 : 124.33Integer char (8bit) compute (GIOPS)char : 126.47char2 : 251.55char4 : 498.03char8 : 497.37char16 : 491.94Integer short (16bit) compute (GIOPS)short : 126.31short2 : 250.90short4 : 249.47short8 : 248.51short16 : 245.30Transfer bandwidth (GBPS)enqueueWriteBuffer : 8.54enqueueReadBuffer : 9.97enqueueWriteBuffer non-blocking : 8.55enqueueReadBuffer non-blocking : 9.99enqueueMapBuffer(for read) : 61.66memcpy from mapped ptr : 11.95enqueueUnmap(after write) : 62.02memcpy to mapped ptr : 11.89Kernel launch latency : 26.81 us
5.使用OpenBLAS测试CPU的算力
git clone https://github.com/xianyi/OpenBLAS.git
cd OpenBLAS
make TARGET=ARMV8
make install
cd benchmark
make TARGET=ARMV8 sgemm
cc sgemm.o -o sgemm /opt/OpenBLAS/lib/libopenblas.so -Wl,-rpath=/opt/OpenBLAS/lib/
export OPENBLAS_NUM_THREADS=8
export OPENBLAS_LOOPS=10
export OPENBLAS_PARAM_M=8192
export OPENBLAS_PARAM_N=8192
export OPENBLAS_PARAM_K=8192
./sgemm
输出
From : 1 To : 200 Step=1 : Transa=N : Transb=NSIZE Flops TimeM=8192, N=8192, K=8192 : 53485.68 MFlops 205.571220 sec
opencv_resize_629">6.分别用CPU与OpenCL测试opencv resize的性能
A.编译OpenCV支持OpenCL
- Opencv修改点[链接libmali.so]
diff --git a/cmake/OpenCVDetectOpenCL.cmake b/cmake/OpenCVDetectOpenCL.cmake
index 6ab2cae070..c3cf235e45 100644
--- a/cmake/OpenCVDetectOpenCL.cmake
+++ b/cmake/OpenCVDetectOpenCL.cmake
@@ -3,9 +3,8 @@ if(APPLE)set(OPENCL_LIBRARY "-framework OpenCL" CACHE STRING "OpenCL library")set(OPENCL_INCLUDE_DIR "" CACHE PATH "OpenCL include directory")else()
- set(OPENCL_LIBRARY "" CACHE STRING "OpenCL library")
- set(OPENCL_INCLUDE_DIR "${OpenCV_SOURCE_DIR}/3rdparty/include/opencl/1.2" CACHE PATH "OpenCL include directory")
- ocv_install_3rdparty_licenses(opencl-headers "${OpenCV_SOURCE_DIR}/3rdparty/include/opencl/LICENSE.txt")
+ set(OPENCL_LIBRARY "/usr/lib/aarch64-linux-gnu/libmali.so")
+ set(OPENCL_INCLUDE_DIR "/usr/include")endif()mark_as_advanced(OPENCL_INCLUDE_DIR OPENCL_LIBRARY)
- 编译Opencv
git clone https://github.com/opencv/opencv.git
cd opencv
git checkout bdb6a968ce69a2bf7c34724f9052c20e941ab47b
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release \-DCMAKE_INSTALL_PREFIX=`pwd`/_install \-DWITH_OPENCL=ON -DWITH_NEON=ON \-DBUILD_SHARED_LIBS=ON \-D BUILD_opencv_world=ON -DBUILD_TESTS=OFF -DBUILD_EXAMPLES=OFF -DBUILD_opencv_apps=OFF \-DBUILD_opencv_dnn=OFF -DBUILD_opencv_calib3d=OFF \-DBUILD_opencv_imgproc=ON -DBUILD_opencv_imgcodecs=ON ..
make -j4
make install
B.运行OpenCV测试程序
cat > opencv_resize.cpp <<-'EOF'
#include <opencv2/opencv.hpp>
#include <opencv2/core/ocl.hpp>
#include <iostream>
#include <map>void run(int resize_mode)
{// 创建一个32x32的随机图像cv::Mat src = cv::Mat::zeros(32, 32, CV_8UC3);cv::randu(src, cv::Scalar::all(0), cv::Scalar::all(255));// ------------------------------------// 在CPU上执行// ------------------------------------cv::ocl::setUseOpenCL(false);cv::Mat enlarged_cpu, resized_back_cpu;// 记录放大操作的开始时间int64 start_time_cpu = cv::getTickCount();for(int i=0;i<100;i++){// 放大到8192x8192cv::resize(src, enlarged_cpu, cv::Size(8192, 8192), 0, 0, resize_mode);// 缩小回32x32cv::resize(enlarged_cpu, resized_back_cpu, cv::Size(32, 32), 0, 0, resize_mode);}// 记录缩小操作的结束时间int64 end_time_cpu = cv::getTickCount();// 计算缩小操作的耗时double time_resize_cpu = (end_time_cpu - start_time_cpu) / cv::getTickFrequency();// ------------------------------------// 在GPU（OpenCL）上执行// ------------------------------------cv::ocl::setUseOpenCL(true);cv::UMat src_umat;src.copyTo(src_umat);cv::UMat enlarged_gpu, resized_back_gpu;// 记录放大操作的开始时间int64 start_time_gpu = cv::getTickCount();for(int i=0;i<100;i++){// 放大到8192x8192cv::resize(src_umat, enlarged_gpu, cv::Size(8192, 8192), 0, 0, resize_mode);// 缩小回32x32cv::resize(enlarged_gpu, resized_back_gpu, cv::Size(32, 32), 0, 0, resize_mode);}// 记录缩小操作的结束时间int64 end_time_gpu = cv::getTickCount();// 计算缩小操作的耗时double time_resize_gpu = (end_time_gpu - start_time_gpu) / cv::getTickFrequency();std::cout <<"CPU耗时(秒):" << time_resize_cpu << " " << "GPU耗时(秒):" << time_resize_gpu << std::endl;
}int main() {// 检查系统是否支持OpenCLif (!cv::ocl::haveOpenCL()) {std::cout << "系统不支持OpenCL。" << std::endl;return -1;}// 输出OpenCL设备信息cv::ocl::Context context;if (!context.create(cv::ocl::Device::TYPE_GPU)) {std::cout << "未找到可用的GPU设备，使用CPU执行。" << std::endl;} else {cv::ocl::Device device = cv::ocl::Device::getDefault();std::cout << "使用的OpenCL设备：" << device.name() << std::endl;}// 定义要测试的插值方法std::vector<int> interpolation_methods = {cv::INTER_NEAREST,cv::INTER_LINEAR,cv::INTER_CUBIC,cv::INTER_AREA,cv::INTER_LANCZOS4};// 插值方法的名称，用于输出结果std::vector<std::string> interpolation_names = {"INTER_NEAREST","INTER_LINEAR","INTER_CUBIC","INTER_AREA","INTER_LANCZOS4"};for (size_t i = 0; i < interpolation_methods.size(); ++i) {int interpolation = interpolation_methods[i];std::string method_name = interpolation_names[i];std::cout << "插值方法：" << method_name << " ";run(interpolation);} return 0;
}
EOF
g++ -o opencv_resize opencv_resize.cpp -I _install/include/opencv4 \_install/lib/libopencv_world.so -Wl,-rpath=_install/lib
export OPENBLAS_NUM_THREADS=8
./opencv_resize
输出
arm_release_ver of this libmali is 'g6p0-01eac0', rk_so_ver is '6'.
使用的OpenCL设备：Mali-LODX r0p0
插值方法：INTER_NEAREST CPU耗时(秒):3.01526 GPU耗时(秒):0.0672681
插值方法：INTER_LINEAR CPU耗时(秒):5.3227 GPU耗时(秒):0.0189366
插值方法：INTER_CUBIC CPU耗时(秒):8.22734 GPU耗时(秒):11.6337
插值方法：INTER_AREA CPU耗时(秒):20.4999 GPU耗时(秒):27.3197
插值方法：INTER_LANCZOS4 CPU耗时(秒):29.3602 GPU耗时(秒):43.9484
