在 Linux 中在多个平台上启用 OpenCL？如何处理 ICD 文件？
Enable OpenCL over multiple platforms in Linux? How to proceed with ICD files?
详情
我必须安装什么 drivers/packages 才能在多个平台上启用 OpenCL：CPU (Intel)、集成 GPU (Intel)、专用 GPU (NVIDIA)？
最好有所有平台 运行 OpenCL 1.2 或更高版本
我知道这可能是一个简单的修复，也许只是 libraries/SKDs 的正确选择，但我在让更多的平台运行时遇到了一些麻烦。
运行 ubuntu 14.04: 我有一个带有集成英特尔显卡的英特尔酷睿 I5 和一块专用的 NVIDIA Geforce 710m 板。
我用过的资源
https://wiki.tiker.net/OpenCLHowTo
在这里（在 Debian 下）它告诉我我只需要：
- ICD加载器包：（你只需要其中一个）
- ICD 包
- 包 headers
我已经尝试过的
- 已安装 CUDA7.5（全部安装）
- 由于驱动冲突导致黑屏
- 通过卸载所有nvidia驱动并安装352解决
- 仍未安装 Cuda SDK
- 发件人：How to make OpenCL work on 14.10 + Nvidia 331.89 drivers?
sudo apt-get install nvidia-331 nvidia-331-uvm nvidia-opencl-dev nvidia-modprobe- 这些软件包将我的驱动程序降级为 331 和 340
- 也来自：How to make OpenCL work on 14.10 + Nvidia 331.89 drivers?
- 链接库有：
sudo ln -s /usr/include/nvidia-352/GL /usr/local/includesudo ln -s /usr/lib/x86_64-linux-gnu/libOpenCL.so.1 /usr/local/lib/libOpenCL.so
- OpenCL 1.1 适用于 NVIDIA GPU
- 无法获取 OpenCL 1.2，因此决定卸载 331 和 340 并重新安装 352
- 已安装 352（再次）
- OpenCL 1.1 停止工作 NVIDIA GPU（并且仍然不工作）
已安装英特尔 opencl_runtime_14.2_x64_4.5.0.8.tgz
为英特尔 ICD 创建了一个符号 link：
sudo ln -s /opt/intel/opencl-1.2-4.5.0.8/etc/intel64.icd
OpenCL 1.2 为 CPU 工作（并且仍然有效）
已安装 clinfo sudo apt-get install clinfo
- 仅检测到英特尔 CPU 平台
- 试图安装几个不同的 NVIDIA 软件包以使 NVIDIA GPU 再次工作，但没有成功
已安装的软件包和一些信息：
供应商中的 ICD?
ls -l /etc/OpenCL/vendors/
total 4
-rw-r--r-- 1 root root 15 Out 22 2015 Altera.icd
lrwxrwxrwx 1 root root 45 Abr 28 13:48 intel64.icd -> /opt/intel/opencl-1.2-4.5.0.8/etc/intel64.icd
注意缺失的nvidia.icd
CL 和 GL - GL 有一个有效的 link... 现在是红色的
ls -l /usr/local/include
total 4
lrwxrwxrwx 1 root root 31 Abr 28 12:48 CL -> /usr/local/cuda-7.5/include/CL/
lrwxrwxrwx 1 root root 26 Abr 27 11:44 GL -> /usr/include/nvidia-352/GL (IN RED COLOR - folder doesn't exist anymore)
.so 文件
ls -l /usr/local/lib/ | grep CL
lrwxrwxrwx 1 root root 40 Abr 27 11:45 libOpenCL.so -> /usr/lib/x86_64-linux-gnu/libOpenCL.so.1
已安装的软件包
dpkg --get-selections | grep nvidia
nvidia-340 deinstall
nvidia-352 install
nvidia-libopencl1-340 deinstall
nvidia-libopencl1-340-updates deinstall
nvidia-libopencl1-352 deinstall
nvidia-libopencl1-352-updates install
nvidia-modprobe install
nvidia-opencl-icd-340 deinstall
nvidia-opencl-icd-352 deinstall
nvidia-prime install
nvidia-settings install
dpkg --get-selections | grep opencl
nvidia-libopencl1-340 deinstall
nvidia-libopencl1-340-updates deinstall
nvidia-libopencl1-352 deinstall
nvidia-libopencl1-352-updates install
nvidia-opencl-icd-340 deinstall
nvidia-opencl-icd-352 deinstall
ocl-icd-libopencl1:amd64 deinstall
ocl-icd-libopencl1:i386 deinstall
opencl-headers install
unity-scope-openclipart install
clinfo
clinfo
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 LINUX
Platform Name: Intel(R) OpenCL
Platform Vendor: Intel(R) Corporation
Platform Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64
Platform Name: Intel(R) OpenCL
Number of devices: 1
Device Type: CL_DEVICE_TYPE_CPU
Device ID: 32902
Max compute units: 4
Max work items dimensions: 3
Max work items[0]: 8192
Max work items[1]: 8192
Max work items[2]: 8192
Max work group size: 8192
Preferred vector width char: 1
Preferred vector width short: 1
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 16
Native vector width short: 8
Native vector width int: 4
Native vector width long: 2
Native vector width float: 8
Native vector width double: 4
Max clock frequency: 1800Mhz
Address bits: 64
Max memory allocation: 2040185856
Image support: Yes
Max number of images read arguments: 480
Max number of images write arguments: 480
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 480
Max size of kernel argument: 3840
Alignment (bits) of base address: 1024
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: Yes
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: No
Round to +ve and infinity: No
IEEE754-2008 fused multiply-add: No
Cache type: Read/Write
Cache line size: 64
Cache size: 262144
Global memory size: 8160743424
Constant buffer size: 131072
Max number of constant args: 480
Local memory type: Global
Local memory size: 32768
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: Yes
Queue properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 0x1659390
Name: Intel(R) Core(TM) i5-3337U CPU @ 1.80GHz
Vendor: Intel(R) Corporation
Device OpenCL C version: OpenCL C 1.2
Driver version: 1.2.0.8
Profile: FULL_PROFILE
Version: OpenCL 1.2 (Build 8)
Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64
所以...
如何让 NVIDIA GPU 也显示为 OPENCL 1.2（或更高版本）平台？集成英特尔显卡怎么样？
AMD 库可以与我的硬件一起使用吗？
为什么大多数 nvidia 软件包都标记为 deinstall?
如前所述，需要三样东西：
来自 https://wiki.tiker.net/OpenCLHowTo
- ICD加载器包：（你只需要其中一个）
- ICD 包
- 包 headers
因此对于 Intel CPU 和 NVIDIA GPU
- ICD 加载器 的软件包：
- ocl-icd-libopencl1
- ICD 包裹
- 已安装Intel OpenCL runtime
- nvidia-opencl-icd-352
- 包 headers
- opencl-headers
但是，要使其正常工作，必须确保 dpkg --get-selections | grep opencl
将软件包标记为已安装
sudo apt-get install --reinstall nvidia-opencl-icd-352 opencl-headers ocl-icd-libopencl1
最重要的是，您必须确保 intel64.icd 和 nvidia.icd 在 /etc/OpenCL/vendors (ls -l /etc/OpenCL/vendors) 中。
也就是说，我必须 link intel64.icd 与：
cd /etc/OpenCL/vendors/
sudo ln -s /opt/intel/opencl-1.2-X.X.X.X/etc/intel64.icd
而且，由于 nvidia.icd 不在文件夹中（即使在安装了正确的包之后）我不得不解压缩它并手动从 deb 包中移动
dpkg -x /var/cache/apt/archives/nvidia-opencl-icd-352_352.63-0ubuntu0.14.04.1_amd64.deb ~/tempfolder
sudo mv ~/tempfolder/etc/OpenCL/vendors/nvidia.icd /etc/OpenCL/vendors/nvidia.icd
rm -r ~/tempfolder
最后，确保 nvidia 是活动 GPU
sudo prime-select nvidia
sudo reboot -r now
安装并执行 clinfo，两个平台都应该 show-up。
clinfo
Number of platforms: 2
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 CUDA 7.5.23
Platform Name: NVIDIA CUDA
Platform Vendor: NVIDIA Corporation
Platform Extensions: cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 LINUX
Platform Name: Intel(R) OpenCL
Platform Vendor: Intel(R) Corporation
Platform Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64
详情
我必须安装什么 drivers/packages 才能在多个平台上启用 OpenCL：CPU (Intel)、集成 GPU (Intel)、专用 GPU (NVIDIA)？
最好有所有平台 运行 OpenCL 1.2 或更高版本
我知道这可能是一个简单的修复，也许只是 libraries/SKDs 的正确选择，但我在让更多的平台运行时遇到了一些麻烦。
运行 ubuntu 14.04: 我有一个带有集成英特尔显卡的英特尔酷睿 I5 和一块专用的 NVIDIA Geforce 710m 板。
我用过的资源
https://wiki.tiker.net/OpenCLHowTo
在这里（在 Debian 下）它告诉我我只需要：
- ICD加载器包：（你只需要其中一个）
- ICD 包
- 包 headers
我已经尝试过的
- 已安装 CUDA7.5（全部安装）
- 由于驱动冲突导致黑屏
- 通过卸载所有nvidia驱动并安装352解决
- 仍未安装 Cuda SDK
- 由于驱动冲突导致黑屏
- 发件人：How to make OpenCL work on 14.10 + Nvidia 331.89 drivers?
sudo apt-get install nvidia-331 nvidia-331-uvm nvidia-opencl-dev nvidia-modprobe- 这些软件包将我的驱动程序降级为 331 和 340
- 也来自：How to make OpenCL work on 14.10 + Nvidia 331.89 drivers?
- 链接库有：
sudo ln -s /usr/include/nvidia-352/GL /usr/local/includesudo ln -s /usr/lib/x86_64-linux-gnu/libOpenCL.so.1 /usr/local/lib/libOpenCL.so
- OpenCL 1.1 适用于 NVIDIA GPU
- 无法获取 OpenCL 1.2，因此决定卸载 331 和 340 并重新安装 352
- 链接库有：
- 已安装 352（再次）
- OpenCL 1.1 停止工作 NVIDIA GPU（并且仍然不工作）
已安装英特尔
opencl_runtime_14.2_x64_4.5.0.8.tgz为英特尔 ICD 创建了一个符号 link：
sudo ln -s /opt/intel/opencl-1.2-4.5.0.8/etc/intel64.icd
OpenCL 1.2 为 CPU 工作（并且仍然有效）
已安装 clinfo
sudo apt-get install clinfo- 仅检测到英特尔 CPU 平台
- 试图安装几个不同的 NVIDIA 软件包以使 NVIDIA GPU 再次工作，但没有成功
已安装的软件包和一些信息：
供应商中的 ICD?
ls -l /etc/OpenCL/vendors/
total 4
-rw-r--r-- 1 root root 15 Out 22 2015 Altera.icd
lrwxrwxrwx 1 root root 45 Abr 28 13:48 intel64.icd -> /opt/intel/opencl-1.2-4.5.0.8/etc/intel64.icd
注意缺失的nvidia.icd
CL 和 GL - GL 有一个有效的 link... 现在是红色的
ls -l /usr/local/include
total 4
lrwxrwxrwx 1 root root 31 Abr 28 12:48 CL -> /usr/local/cuda-7.5/include/CL/
lrwxrwxrwx 1 root root 26 Abr 27 11:44 GL -> /usr/include/nvidia-352/GL (IN RED COLOR - folder doesn't exist anymore)
.so 文件
ls -l /usr/local/lib/ | grep CL
lrwxrwxrwx 1 root root 40 Abr 27 11:45 libOpenCL.so -> /usr/lib/x86_64-linux-gnu/libOpenCL.so.1
已安装的软件包
dpkg --get-selections | grep nvidia
nvidia-340 deinstall
nvidia-352 install
nvidia-libopencl1-340 deinstall
nvidia-libopencl1-340-updates deinstall
nvidia-libopencl1-352 deinstall
nvidia-libopencl1-352-updates install
nvidia-modprobe install
nvidia-opencl-icd-340 deinstall
nvidia-opencl-icd-352 deinstall
nvidia-prime install
nvidia-settings install
dpkg --get-selections | grep opencl
nvidia-libopencl1-340 deinstall
nvidia-libopencl1-340-updates deinstall
nvidia-libopencl1-352 deinstall
nvidia-libopencl1-352-updates install
nvidia-opencl-icd-340 deinstall
nvidia-opencl-icd-352 deinstall
ocl-icd-libopencl1:amd64 deinstall
ocl-icd-libopencl1:i386 deinstall
opencl-headers install
unity-scope-openclipart install
clinfo
clinfo
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 LINUX
Platform Name: Intel(R) OpenCL
Platform Vendor: Intel(R) Corporation
Platform Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64
Platform Name: Intel(R) OpenCL
Number of devices: 1
Device Type: CL_DEVICE_TYPE_CPU
Device ID: 32902
Max compute units: 4
Max work items dimensions: 3
Max work items[0]: 8192
Max work items[1]: 8192
Max work items[2]: 8192
Max work group size: 8192
Preferred vector width char: 1
Preferred vector width short: 1
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 16
Native vector width short: 8
Native vector width int: 4
Native vector width long: 2
Native vector width float: 8
Native vector width double: 4
Max clock frequency: 1800Mhz
Address bits: 64
Max memory allocation: 2040185856
Image support: Yes
Max number of images read arguments: 480
Max number of images write arguments: 480
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 480
Max size of kernel argument: 3840
Alignment (bits) of base address: 1024
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: Yes
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: No
Round to +ve and infinity: No
IEEE754-2008 fused multiply-add: No
Cache type: Read/Write
Cache line size: 64
Cache size: 262144
Global memory size: 8160743424
Constant buffer size: 131072
Max number of constant args: 480
Local memory type: Global
Local memory size: 32768
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: Yes
Queue properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 0x1659390
Name: Intel(R) Core(TM) i5-3337U CPU @ 1.80GHz
Vendor: Intel(R) Corporation
Device OpenCL C version: OpenCL C 1.2
Driver version: 1.2.0.8
Profile: FULL_PROFILE
Version: OpenCL 1.2 (Build 8)
Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64
所以...
如何让 NVIDIA GPU 也显示为 OPENCL 1.2（或更高版本）平台？集成英特尔显卡怎么样？
AMD 库可以与我的硬件一起使用吗？
为什么大多数 nvidia 软件包都标记为 deinstall?
如前所述，需要三样东西：
来自 https://wiki.tiker.net/OpenCLHowTo
- ICD加载器包：（你只需要其中一个）
- ICD 包
- 包 headers
因此对于 Intel CPU 和 NVIDIA GPU
- ICD 加载器 的软件包：
- ocl-icd-libopencl1
- ICD 包裹
- 已安装Intel OpenCL runtime
- nvidia-opencl-icd-352
- 包 headers
- opencl-headers
但是，要使其正常工作，必须确保 dpkg --get-selections | grep opencl
sudo apt-get install --reinstall nvidia-opencl-icd-352 opencl-headers ocl-icd-libopencl1
最重要的是，您必须确保 intel64.icd 和 nvidia.icd 在 /etc/OpenCL/vendors (ls -l /etc/OpenCL/vendors) 中。
也就是说，我必须 link intel64.icd 与：
cd /etc/OpenCL/vendors/
sudo ln -s /opt/intel/opencl-1.2-X.X.X.X/etc/intel64.icd
而且，由于 nvidia.icd 不在文件夹中（即使在安装了正确的包之后）我不得不解压缩它并手动从 deb 包中移动
dpkg -x /var/cache/apt/archives/nvidia-opencl-icd-352_352.63-0ubuntu0.14.04.1_amd64.deb ~/tempfolder
sudo mv ~/tempfolder/etc/OpenCL/vendors/nvidia.icd /etc/OpenCL/vendors/nvidia.icd
rm -r ~/tempfolder
最后，确保 nvidia 是活动 GPU
sudo prime-select nvidia
sudo reboot -r now
安装并执行 clinfo，两个平台都应该 show-up。
clinfo
Number of platforms: 2
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 CUDA 7.5.23
Platform Name: NVIDIA CUDA
Platform Vendor: NVIDIA Corporation
Platform Extensions: cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 LINUX
Platform Name: Intel(R) OpenCL
Platform Vendor: Intel(R) Corporation
Platform Extensions: cl_khr_icd 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_spir cl_intel_exec_by_local_thread cl_khr_depth_images cl_khr_3d_image_writes cl_khr_fp64