Skip to main content
Tutorial

How to set up OpenCL for GPUs on Linux and Docker [Complete Guide]

Learn how to setup OpenCL on Linux and Docker for both NVIDIA and AMD GPUs!

Avimanyu Bandyopadhyay

OpenCL is an abbreviated form for "Open Computing Language". It is a programming language that can be used across diverse platforms, primarily for accelerated computing. Due to its diverse nature of applicabilities across multiple platforms, it is most often referred to as a cross-platform computing language. You can write programs on OpenCL and run them on a variety of devices including CPUs, GPUs, FPGAs and a lot more.

In this guide, I'm going to focus on GPUs alone. I've worked with both NVIDIA and AMD GPUs and I'm going to show you how you can get them operational with OpenCL in the simplest possible way.

Though I've used Ubuntu for the host system, the Docker part is applicable to all other Linux distributions.

Prerequisites

  • NVIDIA/AMD Graphics card
  • Ubuntu Linux 20.04.2 LTS Desktop/Server 64 bit
  • Docker (for application specific usage)

Right on then, let's get to the details!

Setting up OpenCL for NVIDIA GPUs

I'll first show you how to ensure OpenCL works on your main Ubuntu desktop/server. Once that is done, I'll show you how to run Docker containers for the same purpose with the NVIDIA GPU.

Running OpenCL on the host system

On a fresh Ubuntu system, you need to first install the proprietary NVIDIA driver and CUDA. The latter ensures you get the OpenCL framework bundled with it. Finally, install the clinfo program to ensure you have OpenCL properly installed, showing you your NVIDIA GPU's OpenCL specifications in detail. Let's see how:

Use the ubuntu-drivers devices command to fetch the name of your recommended driver:

iborg@iborg-Nitro-AN515-52:~$ ubuntu-drivers devices
== /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 ==
modalias : pci:v000010DEd00001C8Csv00001025sd00001265bc03sc00i00
vendor   : NVIDIA Corporation
model    : GP107M [GeForce GTX 1050 Ti Mobile]
driver   : nvidia-driver-460 - distro non-free recommended
driver   : nvidia-driver-418-server - distro non-free
driver   : nvidia-driver-390 - distro non-free
driver   : nvidia-driver-450-server - distro non-free
driver   : nvidia-driver-465 - distro non-free
driver   : nvidia-driver-460-server - distro non-free
driver   : xserver-xorg-video-nouveau - distro free builtin

Above, note that the recommended driver is nvidia-driver-460.

Install all necessary packages

So let's install the recommended driver along-with CUDA and the clinfo package mentioned earlier in this section:

sudo apt install nvidia-driver-460 nvidia-cuda-toolkit clinfo

After all the above three packages are installed, restart your Ubuntu desktop/server.

Verify your OpenCL configuration

iborg@iborg-Nitro-AN515-52:~$ clinfo
Number of platforms                               1
  Platform Name                                   NVIDIA CUDA
  Platform Vendor                                 NVIDIA Corporation
  Platform Version                                OpenCL 1.2 CUDA 9.1.84
  Platform Profile                                FULL_PROFILE
  Platform 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_fp64 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 cl_nv_create_buffer
  Platform Extensions function suffix             NV

  Platform Name                                   NVIDIA CUDA
Number of devices                                 1
  Device Name                                     GeForce GTX 1050 Ti
  Device Vendor                                   NVIDIA Corporation
  Device Vendor ID                                0x10de
  Device Version                                  OpenCL 1.2 CUDA
  Driver Version                                  390.143
  Device OpenCL C Version                         OpenCL C 1.2 
  Device Type                                     GPU
  Device Topology (NV)                            PCI-E, 01:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               6
  Max clock frequency                             1620MHz
  Compute Capability (NV)                         6.1
  Device Partition                                (core)
    Max number of sub-devices                     1
    Supported partition types                     None
  Max work item dimensions                        3
  Max work item sizes                             1024x1024x64
  Max work group size                             1024
  Preferred work group size multiple              32
  Warp size (NV)                                  32
  Preferred / native vector sizes                 
    char                                                 1 / 1       
    short                                                1 / 1       
    int                                                  1 / 1       
    long                                                 1 / 1       
    half                                                 0 / 0        (n/a)
    float                                                1 / 1       
    double                                               1 / 1        (cl_khr_fp64)
  Half-precision Floating-point support           (n/a)
  Single-precision Floating-point support         (core)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  Yes
  Double-precision Floating-point support         (cl_khr_fp64)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
  Address bits                                    64, Little-Endian
  Global memory size                              4236312576 (3.945GiB)
  Error Correction support                        No
  Max memory allocation                           1059078144 (1010MiB)
  Unified memory for Host and Device              No
  Integrated memory (NV)                          No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       4096 bits (512 bytes)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        98304 (96KiB)
  Global Memory cache line size                   128 bytes
  Image support                                   Yes
    Max number of samplers per kernel             32
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 images
    Max 2D image size                             16384x32768 pixels
    Max 3D image size                             16384x16384x16384 pixels
    Max number of read image args                 256
    Max number of write image args                16
  Local memory type                               Local
  Local memory size                               49152 (48KiB)
  Registers per block (NV)                        65536
  Max number of constant args                     9
  Max constant buffer size                        65536 (64KiB)
  Max size of kernel argument                     4352 (4.25KiB)
  Queue properties                                
    Out-of-order execution                        Yes
    Profiling                                     Yes
  Prefer user sync for interop                    No
  Profiling timer resolution                      1000ns
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Kernel execution timeout (NV)                 Yes
  Concurrent copy and kernel execution (NV)       Yes
    Number of async copy engines                  2
  printf() buffer size                            1048576 (1024KiB)
  Built-in kernels                                
  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_fp64 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 cl_nv_create_buffer

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  NVIDIA CUDA
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [NV]
  clCreateContext(NULL, ...) [default]            Success [NV]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  No platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  No platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  Invalid device type for platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  No platform

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.2.11
  ICD loader Profile                              OpenCL 2.1

Note that here only the platform name is "NVIDIA CUDA". But CUDA and OpenCL are different from each other.

That's it! Now you can run OpenCL applications with your NVIDIA GPU on your host system!

OpenCL on Docker for NVIDIA GPUs

Now that you have OpenCL up and running on your bare-metal system, let's look into how you can get that installed on a Docker container!

Install the NVIDIA Container Runtime

Here, you have to additionally install the nvidia-container-runtime package.

To be able to install it, you must first add the repository details. Make sure you have Curl installed if you haven't already got it on your system.

sudo apt install curl
curl -s -L https://nvidia.github.io/nvidia-container-runtime/gpgkey | \
  sudo apt-key add -
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-container-runtime/$distribution/nvidia-container-runtime.list | \
  sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list
sudo apt update
sudo apt install nvidia-container-runtime

Creating the Dockerfile

It is necessary that you replicate all that you did on the host system on a fresh new image so that you can use it to launch our custom OpenCL applications on a container (will elaborate on that later).

Create a new directory for your NVIDIA GPU OpenCL project and move into it:

mkdir nvidia-opencl
cd nvidia-opencl

Use your favorite text editor (Vim/Nano or any other) to create the following Dockerfile and save it:

FROM ubuntu:20.04
ARG DEBIAN_FRONTEND=noninteractive
RUN apt-get update && apt-get -y upgrade \
  && apt-get install -y \
    apt-utils \
    unzip \
    tar \
    curl \
    xz-utils \
    ocl-icd-libopencl1 \
    opencl-headers \
    clinfo \
    ;

RUN mkdir -p /etc/OpenCL/vendors && \
    echo "libnvidia-opencl.so.1" > /etc/OpenCL/vendors/nvidia.icd
ENV NVIDIA_VISIBLE_DEVICES all
ENV NVIDIA_DRIVER_CAPABILITIES compute,utility

Building the Dockerfile

So now that you have the necessary Dockerfile to get started, let's build it. I'm naming the image as nvidia-opencl:

docker build -t nvidia-opencl .

Launch the OpenCL Container

Based on the new image that you just built, it's time to launch the new OpenCL container!

First, permit your Linux username on the local machine to connect to the X windows display with the following command:

xhost +local:username

With the following command, you can now directly enter the local container's shell based on the new image just created:

docker run --rm -it --gpus all -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY nvidia-opencl

Verify your OpenCL configuration on Docker

Now that you are inside the container shell, you can run the clinfo command to verify your OpenCL configuration just like you did on the bare-metal host system:

root@7b39b04c019f:/# clinfo
Number of platforms                               1
  Platform Name                                   NVIDIA CUDA
  Platform Vendor                                 NVIDIA Corporation
  Platform Version                                OpenCL 1.2 CUDA 9.1.84
  Platform Profile                                FULL_PROFILE
  Platform 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_fp64 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 cl_nv_create_buffer
  Platform Extensions function suffix             NV

  Platform Name                                   NVIDIA CUDA
Number of devices                                 1
  Device Name                                     GeForce GTX 1050 Ti
  Device Vendor                                   NVIDIA Corporation
  Device Vendor ID                                0x10de
  Device Version                                  OpenCL 1.2 CUDA
  Driver Version                                  390.143
  Device OpenCL C Version                         OpenCL C 1.2 
  Device Type                                     GPU
  Device Topology (NV)                            PCI-E, 01:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               6
  Max clock frequency                             1620MHz
  Compute Capability (NV)                         6.1
  Device Partition                                (core)
    Max number of sub-devices                     1
    Supported partition types                     None
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             1024x1024x64
  Max work group size                             1024
  Preferred work group size multiple              32
  Warp size (NV)                                  32
  Preferred / native vector sizes                 
    char                                                 1 / 1       
    short                                                1 / 1       
    int                                                  1 / 1       
    long                                                 1 / 1       
    half                                                 0 / 0        (n/a)
    float                                                1 / 1       
    double                                               1 / 1        (cl_khr_fp64)
  Half-precision Floating-point support           (n/a)
  Single-precision Floating-point support         (core)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  Yes
  Double-precision Floating-point support         (cl_khr_fp64)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
  Address bits                                    64, Little-Endian
  Global memory size                              4236312576 (3.945GiB)
  Error Correction support                        No
  Max memory allocation                           1059078144 (1010MiB)
  Unified memory for Host and Device              No
  Integrated memory (NV)                          No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       4096 bits (512 bytes)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        98304 (96KiB)
  Global Memory cache line size                   128 bytes
  Image support                                   Yes
    Max number of samplers per kernel             32
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 images
    Max 2D image size                             16384x32768 pixels
    Max 3D image size                             16384x16384x16384 pixels
    Max number of read image args                 256
    Max number of write image args                16
  Local memory type                               Local
  Local memory size                               49152 (48KiB)
  Registers per block (NV)                        65536
  Max number of constant args                     9
  Max constant buffer size                        65536 (64KiB)
  Max size of kernel argument                     4352 (4.25KiB)
  Queue properties                                
    Out-of-order execution                        Yes
    Profiling                                     Yes
  Prefer user sync for interop                    No
  Profiling timer resolution                      1000ns
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Kernel execution timeout (NV)                 Yes
  Concurrent copy and kernel execution (NV)       Yes
    Number of async copy engines                  2
  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_fp64 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 cl_nv_create_buffer

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  NVIDIA CUDA
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [NV]
  clCreateContext(NULL, ...) [default]            Success [NV]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  No platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  No platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  Invalid device type for platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  No platform

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.2.11
  ICD loader Profile                              OpenCL 2.1
root@7b39b04c019f:/#

What does this mean? This means now you can run any OpenCL application from inside this container! You'd only have to remodify the Dockerfile and that will be it.

You can also work with Python applications that require an OpenCL backend. Check my earlier coverage that can very much serve as handy companion to this article. You might want to check it out and play with the Dockerfiles.

How to Dockerize Python Applications With Miniconda [A Hybrid Approach]
Take isolation to the next level by using Miniconda to build your Python based docker images for your application in a hybrid manner.

Setting up OpenCL for AMD GPUs

I'll first show you how to ensure OpenCL works on your main Ubuntu desktop/server. Once that is done, I'll show you how to run Docker containers for the same purpose with the AMD GPU.

Running OpenCL on the host system

On a fresh Ubuntu system, you need to download the "AMDGPU drivers" first from the AMD support page. For a future-proof configuration, you only require installing OpenCL both for legacy and newer AMD GPUs after you have obtained the installation archive (tar.xz).

Finally, install the clinfo program to ensure you have OpenCL properly installed, showing you your AMD GPU's OpenCL specifications in detail. But the entire process can be a bit trickier than expected. Let's see how.

Download the AMDGPU Drivers with Curl

Navigate through the AMD support page and download the relevant driver with Curl. Make sure you have Curl installed.

sudo apt install curl
curl -e https://drivers.amd.com/drivers/linux -O https://drivers.amd.com/drivers/linux/amdgpu-pro-21.10-1247438-ubuntu-20.04.tar.xz

Installation, anomalies and their workarounds

Extract the archive:

tar -Jxvf amdgpu-pro-21.10-1247438-ubuntu-20.04.tar.xz

Move into the new directory:

cd amdgpu-pro-21.10-1247438-ubuntu-20.04

Now, I'm going to install OpenCL both for legacy and newer GPUs:

./amdgpu-install --opencl=legacy,rocr --headless --no-dkms

For a complete overview on its usage, you can use the command ./amdgpu-install -h to learn about how the script works fundamentally. It's similar to a man entry for the command. The --headless option specifies only OpenCL support and --no-dkms tells it NOT to install the amdgpu-dkms and the amdgpu-dkms-firmware packages into the kernel. You don't need that.

For quite some time, it has been found that even though you specify the --no-dkms option, the script doesn't bother to comply and proceeds with installing those unnecessary packages. Adding more to it, if I allowed amdgpu-dkms to install and modify the kernel configuration, the system would refuse to reboot or shutdown thereafter! This happened after I received a kernel update from Ubuntu repositories.

In such a case, here's what I did:

I manually installed the following packages using dpkg -i package-name.deb, present inside the extracted directory:

amdgpu-pin_21.10-1247438_all.deb
amdgpu-core_21.10-1247438_all.deb
amdgpu-pro-core_21.10-1247438_all.deb
libdrm-amdgpu-common_1.0.0-1247438_all.deb
libdrm2-amdgpu_2.4.100-1247438_amd64.deb
libdrm-amdgpu-amdgpu1_2.4.100-1247438_amd64.deb
hsakmt-roct-amdgpu_1.0.9-1247438_amd64.deb
hsa-runtime-rocr-amdgpu_1.3.0-1247438_amd64.deb
comgr-amdgpu-pro_2.0.0-1247438_amd64.deb
hip-rocr-amdgpu-pro_21.10-1247438_amd64.deb
ocl-icd-libopencl1-amdgpu-pro_21.10-1247438_amd64.deb
clinfo-amdgpu-pro_21.10-1247438_amd64.deb
opencl-rocr-amdgpu-pro_21.10-1247438_amd64.deb
libllvm11.0-amdgpu_11.0-1247438_amd64.deb

This made sure that amdgpu-dkms and amdgpu-dkms-firmware could be avoided and leave the kernel untouched. Also, note that I've downloaded the older 21.10 driver even though the newer and latest 21.30 version is available. The reason being the latter refuses to recognize my Radeon VII GPU by giving an "HSA Error" when I run clinfo later:

HSA Error: Incompatible kernel and userspace, Vega 20 [Radeon VII] disabled. Upgrade amdgpu.

After taking care of these anomalies, I was able to get clinfo to report my GPU correctly.

Install the clinfo package

Install the clinfo package like you did earlier for NVIDIA GPUs:

sudo apt install clinfo

Verify your OpenCL configuration

avimanyu@GizmoQuest-Computing-Lab:~$ clinfo
Number of platforms                               1
  Platform Name                                   AMD Accelerated Parallel Processing
  Platform Vendor                                 Advanced Micro Devices, Inc.
  Platform Version                                OpenCL 2.0 AMD-APP (3246.0)
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd cl_amd_event_callback 
  Platform Extensions function suffix             AMD

  Platform Name                                   AMD Accelerated Parallel Processing
Number of devices                                 1
  Device Name                                     gfx906:sramecc-:xnack-
  Device Vendor                                   Advanced Micro Devices, Inc.
  Device Vendor ID                                0x1002
  Device Version                                  OpenCL 2.0 
  Driver Version                                  3246.0 (HSA1.1,LC)
  Device OpenCL C Version                         OpenCL C 2.0 
  Device Type                                     GPU
  Device Board Name (AMD)                         Vega 20 [Radeon VII]
  Device Topology (AMD)                           PCI-E, 0a:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               60
  SIMD per compute unit (AMD)                     4
  SIMD width (AMD)                                16
  SIMD instruction width (AMD)                    1
  Max clock frequency                             1801MHz
  Graphics IP (AMD)                               9.0
  Device Partition                                (core)
    Max number of sub-devices                     60
    Supported partition types                     None
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             1024x1024x1024
  Max work group size                             256
  Preferred work group size (AMD)                 256
  Max work group size (AMD)                       1024
  Preferred work group size multiple              64
  Wavefront width (AMD)                           64
  Preferred / native vector sizes                 
    char                                                 4 / 4       
    short                                                2 / 2       
    int                                                  1 / 1       
    long                                                 1 / 1       
    half                                                 1 / 1        (cl_khr_fp16)
    float                                                1 / 1       
    double                                               1 / 1        (cl_khr_fp64)
  Half-precision Floating-point support           (cl_khr_fp16)
    Denormals                                     No
    Infinity and NANs                             No
    Round to nearest                              No
    Round to zero                                 No
    Round to infinity                             No
    IEEE754-2008 fused multiply-add               No
    Support is emulated in software               No
  Single-precision Floating-point support         (core)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  Yes
  Double-precision Floating-point support         (cl_khr_fp64)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
  Address bits                                    64, Little-Endian
  Global memory size                              17163091968 (15.98GiB)
  Global free memory (AMD)                        16760832 (15.98GiB)
  Global memory channels (AMD)                    128
  Global memory banks per channel (AMD)           4
  Global memory bank width (AMD)                  256 bytes
  Error Correction support                        No
  Max memory allocation                           14588628168 (13.59GiB)
  Unified memory for Host and Device              No
  Shared Virtual Memory (SVM) capabilities        (core)
    Coarse-grained buffer sharing                 Yes
    Fine-grained buffer sharing                   Yes
    Fine-grained system sharing                   No
    Atomics                                       No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       1024 bits (128 bytes)
  Preferred alignment for atomics                 
    SVM                                           0 bytes
    Global                                        0 bytes
    Local                                         0 bytes
  Max size for global variable                    14588628168 (13.59GiB)
  Preferred total size of global vars             17163091968 (15.98GiB)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        16384 (16KiB)
  Global Memory cache line size                   64 bytes
  Image support                                   Yes
    Max number of samplers per kernel             26287
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 8192 images
    Base address alignment for 2D image buffers   256 bytes
    Pitch alignment for 2D image buffers          256 pixels
    Max 2D image size                             16384x16384 pixels
    Max 3D image size                             16384x16384x8192 pixels
    Max number of read image args                 128
    Max number of write image args                8
    Max number of read/write image args           64
  Max number of pipe args                         16
  Max active pipe reservations                    16
  Max pipe packet size                            1703726280 (1.587GiB)
  Local memory type                               Local
  Local memory size                               65536 (64KiB)
  Local memory syze per CU (AMD)                  65536 (64KiB)
  Local memory banks (AMD)                        32
  Max number of constant args                     8
  Max constant buffer size                        14588628168 (13.59GiB)
  Preferred constant buffer size (AMD)            16384 (16KiB)
  Max size of kernel argument                     1024
  Queue properties (on host)                      
    Out-of-order execution                        No
    Profiling                                     Yes
  Queue properties (on device)                    
    Out-of-order execution                        Yes
    Profiling                                     Yes
    Preferred size                                262144 (256KiB)
    Max size                                      8388608 (8MiB)
  Max queues on device                            1
  Max events on device                            1024
  Prefer user sync for interop                    Yes
  Number of P2P devices (AMD)                     0
  P2P devices (AMD)                               <printDeviceInfo:147: get number of CL_DEVICE_P2P_DEVICES_AMD : error -30>
  Profiling timer resolution                      1ns
  Profiling timer offset since Epoch (AMD)        0ns (Thu Jan  1 05:30:00 1970)
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Thread trace supported (AMD)                  No
    Number of async queues (AMD)                  8
    Max real-time compute queues (AMD)            8
    Max real-time compute units (AMD)             60
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_fp64 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_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_media_ops cl_amd_media_ops2 cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_depth_images cl_amd_copy_buffer_p2p cl_amd_assembly_program 

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  No platform
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   No platform
  clCreateContext(NULL, ...) [default]            No platform
  clCreateContext(NULL, ...) [other]              Success [AMD]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-

So, now you can run OpenCL applications with your AMD GPU on your host system!

OpenCL on Docker for AMD GPUs

How about doing the same through Docker containers? Let's see how much it contrasts with NVIDIA GPUs.

Creating the Dockerfile

Create a new directory for your AMD GPU OpenCL project and move into it:

mkdir amd-opencl
cd amd-opencl

Use your favorite text editor (Vim/Nano or any other) to create the following Dockerfile and save it:

FROM ubuntu:20.04
ARG DEBIAN_FRONTEND=noninteractive
RUN apt-get update && apt-get -y upgrade \
  && apt-get install -y \
    initramfs-tools \
    apt-utils \
    unzip \
    tar \
    curl \
    xz-utils \
    ocl-icd-libopencl1 \
    opencl-headers \
    clinfo \
    ;

ARG AMD_DRIVER=amdgpu-pro-21.10-1247438-ubuntu-20.04.tar.xz
ARG AMD_DRIVER_URL=https://drivers.amd.com/drivers/linux
RUN mkdir -p /tmp/opencl-driver-amd
WORKDIR /tmp/opencl-driver-amd
RUN curl --referer $AMD_DRIVER_URL -O $AMD_DRIVER_URL/$AMD_DRIVER; \
    tar -Jxvf $AMD_DRIVER; \
    cd amdgpu-pro-*; \
    ./amdgpu-install --opencl=legacy,rocr --headless --no-dkms -y; \
    rm -rf /tmp/opencl-driver-amd;

RUN mkdir -p /etc/OpenCL/vendors && \
    echo "libamdocl64.so" > /etc/OpenCL/vendors/amdocl64.icd
RUN ln -s /usr/lib/x86_64-linux-gnu/libOpenCL.so.1 /usr/lib/libOpenCL.so
WORKDIR /

I had to add the initramfs-tools package since the amdgpu-dkms and amdgpu-dkms-firmware would still be installed. I kept it this way since in this case, the reboot and shutdown issues I mentioned earlier are irrelevant for containers.

Alternatively, you could still use the dpkg -i method in the Dockerfile.

Building the Dockerfile

So now that you have the necessary Dockerfile to get started, let's build it. I'm naming the image as amd-opencl:

docker build -t amd-opencl .

Add your username to the video & render groups

For the AMD GPU Docker container to work flawlessly, it is better you also add your username to the video and render groups:

sudo usermod -a -G video $LOGNAME
sudo usermod -a -G render $LOGNAME

Launch the OpenCL Container

Based on the new image that you just built, it's time to launch the new OpenCL container!

Permit your Linux username on the local machine to connect to the X windows display with the following command:

xhost +local:username

With the following command, you can now directly enter the local container's shell based on the new image just created:

docker run --rm -it --device=/dev/kfd --device=/dev/dri --group-add video --group-add render -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY amd-opencl

Verify your OpenCL configuration on Docker

Now that you are inside the container shell, you can run the clinfo command to verify your OpenCL configuration just like you did on the bare-metal host system:

root@00ec73e147bc:/# clinfo
Platform Name                                   AMD Accelerated Parallel Processing
Number of devices                                 1
  Device Name                                     gfx906:sramecc-:xnack-
  Device Vendor                                   Advanced Micro Devices, Inc.
  Device Vendor ID                                0x1002
  Device Version                                  OpenCL 2.0 
  Driver Version                                  3246.0 (HSA1.1,LC)
  Device OpenCL C Version                         OpenCL C 2.0 
  Device Type                                     GPU
  Device Board Name (AMD)                         Device 66af
  Device Topology (AMD)                           PCI-E, 0a:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               60
  SIMD per compute unit (AMD)                     4
  SIMD width (AMD)                                16
  SIMD instruction width (AMD)                    1
  Max clock frequency                             1801MHz
  Graphics IP (AMD)                               9.0
  Device Partition                                (core)
    Max number of sub-devices                     60
    Supported partition types                     None
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             1024x1024x1024
  Max work group size                             256
  Preferred work group size (AMD)                 256
  Max work group size (AMD)                       1024
  Preferred work group size multiple              64
  Wavefront width (AMD)                           64
  Preferred / native vector sizes                 
    char                                                 4 / 4       
    short                                                2 / 2       
    int                                                  1 / 1       
    long                                                 1 / 1       
    half                                                 1 / 1        (cl_khr_fp16)
    float                                                1 / 1       
    double                                               1 / 1        (cl_khr_fp64)
  Half-precision Floating-point support           (cl_khr_fp16)
    Denormals                                     No
    Infinity and NANs                             No
    Round to nearest                              No
    Round to zero                                 No
    Round to infinity                             No
    IEEE754-2008 fused multiply-add               No
    Support is emulated in software               No
  Single-precision Floating-point support         (core)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  Yes
  Double-precision Floating-point support         (cl_khr_fp64)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
  Address bits                                    64, Little-Endian
  Global memory size                              17163091968 (15.98GiB)
  Global free memory (AMD)                        16760832 (15.98GiB)
  Global memory channels (AMD)                    128
  Global memory banks per channel (AMD)           4
  Global memory bank width (AMD)                  256 bytes
  Error Correction support                        No
  Max memory allocation                           14588628168 (13.59GiB)
  Unified memory for Host and Device              No
  Shared Virtual Memory (SVM) capabilities        (core)
    Coarse-grained buffer sharing                 Yes
    Fine-grained buffer sharing                   Yes
    Fine-grained system sharing                   No
    Atomics                                       No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       1024 bits (128 bytes)
  Preferred alignment for atomics                 
    SVM                                           0 bytes
    Global                                        0 bytes
    Local                                         0 bytes
  Max size for global variable                    14588628168 (13.59GiB)
  Preferred total size of global vars             17163091968 (15.98GiB)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        16384 (16KiB)
  Global Memory cache line size                   64 bytes
  Image support                                   Yes
    Max number of samplers per kernel             26287
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 8192 images
    Base address alignment for 2D image buffers   256 bytes
    Pitch alignment for 2D image buffers          256 pixels
    Max 2D image size                             16384x16384 pixels
    Max 3D image size                             16384x16384x8192 pixels
    Max number of read image args                 128
    Max number of write image args                8
    Max number of read/write image args           64
  Max number of pipe args                         16
  Max active pipe reservations                    16
  Max pipe packet size                            1703726280 (1.587GiB)
  Local memory type                               Local
  Local memory size                               65536 (64KiB)
  Local memory syze per CU (AMD)                  65536 (64KiB)
  Local memory banks (AMD)                        32
  Max number of constant args                     8
  Max constant buffer size                        14588628168 (13.59GiB)
  Preferred constant buffer size (AMD)            16384 (16KiB)
  Max size of kernel argument                     1024
  Queue properties (on host)                      
    Out-of-order execution                        No
    Profiling                                     Yes
  Queue properties (on device)                    
    Out-of-order execution                        Yes
    Profiling                                     Yes
    Preferred size                                262144 (256KiB)
    Max size                                      8388608 (8MiB)
  Max queues on device                            1
  Max events on device                            1024
  Prefer user sync for interop                    Yes
  Number of P2P devices (AMD)                     0
  P2P devices (AMD)                               <printDeviceInfo:147: get number of CL_DEVICE_P2P_DEVICES_AMD : error -30>
  Profiling timer resolution                      1ns
  Profiling timer offset since Epoch (AMD)        0ns (Thu Jan  1 00:00:00 1970)
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Thread trace supported (AMD)                  No
    Number of async queues (AMD)                  8
    Max real-time compute queues (AMD)            8
    Max real-time compute units (AMD)             60
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_fp64 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_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_media_ops cl_amd_media_ops2 cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_depth_images cl_amd_copy_buffer_p2p cl_amd_assembly_program

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  No platform
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   No platform
  clCreateContext(NULL, ...) [default]            No platform
  clCreateContext(NULL, ...) [other]              Success [AMD]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  Success (1)
    Platform Name                                 AMD Accelerated Parallel Processing
    Device Name                                   gfx906:sramecc-:xnack-
root@00ec73e147bc:/#

And that's how you can run OpenCL applications inside an AMD GPU container!

Note that the xhost command being used for both the NVIDIA and AMD GPU containers is necessary every time you want to run them from a new terminal.

Bonus Tips

If you happen to own multiple GPUs on a single system and want to be specific about running the containers, you can do that as well. Read on.

NVIDIA GPUs

Based on how clinfo reports NVIDIA GPU information, they are classified on Docker as 0, 1 , 2 and so on. So, say you have three NVIDIA GPUs and want the container to see only GPU 0(the first one), the corresponding command would have to be revised as:

docker run --rm -it --gpus 0 -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY nvidia-opencl

AMD GPUs

Similarly, based on how clinfo reports AMD GPU information, they are classified on Docker as /dev/dri/card0, /dev/dri/card1, /dev/dri/card2 and so on. So, say you have three AMD GPUs and want the container to see only the first, use the following command instead:

docker run --rm -it --device=/dev/kfd --device=/dev/dri/card0 --device=/dev/dri/renderD128 --group-add video --group-add render -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY amd-opencl

As per the above command, note that renderD128 corresponds to card0, both of which relate to the first AMD GPU. On the same lines, renderD129 would correspond to card1 for the second AMD GPU and so on. The "renderD" value is incremental and therefore for the third GPU, it would be renderD130 corresponding to card2. You can know these metrics in detail by running the ls -l /dev/dri/by-path command.

Personal notes

Since the last 7 years, I've been actively involved with research that focuses on harnessing the computational power of Graphics Processing Units (GPUs) to understand biological phenomena.

For more than a year now, I've been working on Dockerizing CellModeller, which is my primary research software that I've been working with, to understand multicellularity and at the same time also contributing on its development as a software.

Even though the AMD GPU containerization process can be a bit tedious and tricky, I still liked the way it works without the need of an additional runtime package necessary for NVIDIA GPU containers.

For the entire endeavour, the following references were extremely helpful:

Congleton, N., 2020. Install OpenCL For The AMDGPU Open Source Drivers On Debian and Ubuntu. [online] LinuxConfig.org. Available at:  https://linuxconfig.org/install-opencl-for-the-amdgpu-open-source-drivers-on-debian-and-ubuntu [Accessed June 23 2021].
amdgpu and OpenCL
pkienzle/opencl_docker
yet another opencl docker container (amd, nvidia, intel cpu) - pkienzle/opencl_docker

My heartfelt thanks to all three authors!

There are so many applications out there on the accelerated computing domain that need OpenCL running on the backend for both GPU vendors. One good example is Folding@home and its specific GPU requirements.

Do let me know your thoughts about this intriguing adventure with OpenCL, GPUs, Linux and finally, Docker! If you have any feedback or suggestions, please let me know in the comment section below.

Avimanyu Bandyopadhyay
Website @iAvimanyu Facebook Kolkata, West Bengal, India