logo Debian Debian Debian-France Debian-Facile Debian-fr.org Forum-Debian.fr Debian ? Communautés logo inclusivité

Debian-facile

Bienvenue sur Debian-Facile, site d'aide pour les nouveaux utilisateurs de Debian.

Vous n'êtes pas identifié(e).


L'icône rouge permet de télécharger chaque page du wiki visitée au format PDF et la grise au format ODT → ODT PDF Export

Ceci est une ancienne révision du document !


openCL avec carte graphique AMD

Nota :

Contributeurs, les FIXME sont là pour vous aider, supprimez-les une fois le problème corrigé ou le champ rempli !

Introduction

Ce petit tuto reprend les étapes d'installation du pilote openCL sur ces matériels et système :

  • carte AMD RX 550 4GB
  • carte AMD RX 570 8GB ITX

le tout , fait sur linux debian testing bullseye et avec

  • X.org GNOME 3.34.2 et 3.36.2
  • blender 2.82a
  • processeur AMD RYZEN 7 1700 X

Un grand merci à PLOP6 via (irc) sans qui rien de ce tuto n'aurait été possible.

Historique

Le pilote OpenCL, a été créé pour uniformiser “massivement” les programmes de calculs parallèles2).
Au départ , son emploi s'étendait aussi bien aux calculs fait par les CPU que ceux faits par les GPU, possibilité que l'on retrouve aujourd'hui dans Blender.

  • OpenCL donne la possibilité d'utiliser ou/et le CPU ou/et le GPU.
  • OpenCL est de plus en plus utilisé pour employer la puissance de calcul des cartes graphiques.

Petite bibliographie openCL

Petite bibliographie dtrx

Préambule

Un conseil : travaillez dans le répertoire ~/Téléchargements , c'est plus sûr.

Téléchargements

Avec la commande cd, se positionner dans le répertoire Téléchargements ainsi :

cd ~/Téléchargements/

Téléchargement de l'archive du pilote amdgpu-pro

  1. Soit : (version 19.50 pour ubuntu 18.04 LTS) :
    wget --referer=https://www.amd.com/ -P $HOME/Téléchargements/ https://drivers.amd.com/drivers/linux/19.50/amdgpu-pro-19.50-967956-ubuntu-18.04.tar.xz
  2. Soit : (version 20.10 pour ubuntu 18.04 LTS) :
     wget --referer=https://www.amd.com/ -P $HOME/Téléchargements/ https://drivers.amd.com/drivers/linux/amdgpu-pro-20.10-1048554-ubuntu-18.04.tar.xz

Au cas où ces liens ne fonctionnent pas, vous pouvez vous connecter en graphique là :

  1. Cherchez la ligne
    + Ubuntu X86 64-bit

    et trouvez la case Téléchargement sur laquelle vous cliquez.

  2. Vous téléchargez ensuite l'archive dans le répertoire ~/Téléchargements de votre Debian.
sudo a l'avantage de passer des commandes root tout en restant user
  1. Installer dtrx3) :
    sudo apt install dtrx 
  2. Si cela ne fonctionne pas4) :
    pip install dtrx
  3. installer aussi clinfo et mesa-utils :
    sudo apt install clinfo mesa-utils 

Scripts

Créer les scripts suivants :

copiez/collez chacun des blocs ci-dessous et lancez-les indépendamment en user, chacun d'eux créera son fichier dans “ ~/Téléchargements/ ”

 cat > ~/Téléchargements/Arborescence.sh <<EOF
 
#!/bin/bash
 
mkdir -p /opt/amdgpu-pro/lib/x86_64-linux-gnu
mkdir -p /etc/OpenCL/vendors
 
EOF
cat > ~/Téléchargements/Extraction.sh <<EOF
 
#!/bin/bash
 
if ! which dtrx > /dev/null; then
   echo -e "le paquet dtrx n'est pas installé. Install it and re-run script!"
   echo "Error No PAckage dtrx" >> error_OnExtraitToutD1coup
   exit 1
fi
dtrx -r -q amdgpu-pro-*.tar.xz
 
 
EOF
le script ci-dessous5) est souvent capricieux, assurez-vous bien de sa copie avant de le lancer, quitte à le copier à la main6) avec nano .
cat > ~/Téléchargements/Copie.sh <<EOF
 
#!/bin/bash
 
LibDir="/opt/amdgpu-pro/lib/x86_64-linux-gnu/"
VendorsDir="/etc/OpenCL/vendors/"
DriversDir="amdgpu-pro-*/"
 
# on Copie les libs orca:
CustDir="${DriversDir}opencl-orca-amdgpu-pro-icd_*/opt/amdgpu-pro/lib/x86_64-linux-gnu/"
for libso in libamdocl12cl64.so libamdocl-orca64.so; do
    cp ${CustDir}${libso} ${LibDir}
done
 
# on Copie les libs Opencl:
CustDir="${DriversDir}libopencl1-amdgpu-pro_*/opt/amdgpu-pro/lib/x86_64-linux-gnu/"
for libso in  libOpenCL.so libOpenCL.so.1;do
    cp ${CustDir}${libso} ${LibDir}
done
 
# On copie libcltrace
CustDir="${DriversDir}opencl-amdgpu-pro-dev_*/opt/amdgpu-pro/lib/x86_64-linux-gnu/"
cp ${CustDir}libcltrace.so ${LibDir}
 
# On Copie libamdocl64
CustDir="${DriversDir}opencl-amdgpu-pro-icd_*/opt/amdgpu-pro/lib/x86_64-linux-gnu/"
cp ${CustDir}libamdocl64.so ${LibDir}
 
# On copie les Fichiers ICD
CustDir="${DriversDir}opencl-orca-amdgpu-pro-icd_*/etc/OpenCL/vendors/"
cp ${CustDir}amdocl-orca64.icd ${VendorsDir}
CustDir="${DriversDir}opencl-amdgpu-pro-icd_*/etc/OpenCL/vendors/"
cp ${CustDir}amdocl64.icd ${VendorsDir}
 
EOF
 cat > ~/Téléchargements/maj.sh <<EOF
 
#!/bin/bash
 
echo "/opt/amdgpu-pro/lib/x86_64-linux-gnu" > /etc/ld.so.conf.d/amdgpu-pro_custom-opencl.conf
 
ldconfig
 
EOF
cat > ~/Téléchargements/raz.sh <<EOF
 
#!/bin/bash
 
rm -rf /etc/OpenCL
rm -rf /opt/amdgpu-pro
rm /etc/ld.so.conf.d/amdgpu-pro_custom-opencl.conf
ldconfig
 
EOF
Une fois les scripts copiés/collés, enregistrés, pour pouvoir les utiliser en user7), n'oubliez pas de modifier les droits d'utilisations des scripts8) avec la commande chmod.

Autorisation des scripts (les rendre exécutables)

puis on donne les autorisations au fichier :

sudo chmod u+x ~/Téléchargements/Arborescence.sh
sudo chmod u+x ~/Téléchargements/Extraction.sh
sudo chmod u+x ~/Téléchargements/Copie.sh
sudo chmod u+x ~/Téléchargements/maj.sh
sudo chmod u+x ~/Téléchargements/raz.sh

Installation

commencer par se placer dans le répertoire Téléchargements :
cd ~/Téléchargements/

On commence par créer l'arborescence : -

sudo bash ~/Téléchargements/Arborescence.sh 
  1. on décompacte les deb :
    sudo bash ~/Téléchargements/Extraction.sh 
  2. on copie les bons fichiers :
  3. sudo bash ~/Téléchargements/Copie.sh
  4. on met à jour :
  5. sudo bash ~/Téléchargements/maj.sh
  6. en cas de doute ou de besoin , on peut tout virer :
    sudo bash ~/Téléchargements/raz.sh 

Script git

un script git qui télécharge amdgpu-pro et installe OpenCL et Vulkan .

actuellement la version 21.30 du 04/08/2021 pour ubuntu 20.04.3 LTS . fonctionne, en théorie , pour RX 5000 ET RX 6000 . pas de prise en charge du noyau 5.11 (en théorie). toujours non fonctionnel sur debian sid . (noyau 5.10)

-- script git --

git clone https://gist.github.com/3351b5d0b4f947e19df36b1ea3c95cbe.git
installer “git” au préalable
sudo apt install --reinstall git 
merci à kytulendu

Utilisation

pour tester blender avec openCL , on peut télécharger des exemples ici :
exemples blender

dans Edit / Preferences / System / OpenCL ,

vous pourrez choisir votre carte graphique (ici AMD Radeon RX 550 / 550 Series) et / ou votre processeur (AMD RYZEN 7 1700X , par exemple)

un grand merci à plop6 (irc) sans qui rien n'aurai été possible …

Post Scriptum

la carte (RX 550) est montée en pointe , à 75 °C , au maxi une conso de 22 watts et un ventilo de 1300 RPM . la RX 570 , elle , est montée , en pointe à 66 °C , conso maxi 100 watts et un ventilo à 1900 Tr/Min .

voici quelques écrans que vous obtiendrez peut être à la fin de l'installation :

tests fait avec la version 19.50 .

la version 20.10 donne le même résultat . (légèrement updaté)

AMD RX 550 4GB DDR5

ma carte graphique :

lspci -nnkd::0300
0a:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Lexa PRO [Radeon 540/540X/550/550X / RX 540X/550/550X] [1002:699f] (rev c7)
	Subsystem: Sapphire Technology Limited Lexa PRO [Radeon RX 550] [1da2:e367]
	Kernel driver in use: amdgpu
	Kernel modules: amdgpu

son pilote openGL (amdgpu) :

glxinfo -B
name of display: :0
display: :0  screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
    Vendor: X.Org (0x1002)
    Device: Radeon RX550/550 Series (POLARIS12, DRM 3.35.0, 5.4.0-4-amd64, LLVM 9.0.1) (0x699f)
    Version: 19.3.3
    Accelerated: yes
    Video memory: 4096MB
    Unified memory: no
    Preferred profile: core (0x1)
    Max core profile version: 4.5
    Max compat profile version: 4.5
    Max GLES1 profile version: 1.1
    Max GLES[23] profile version: 3.2
Memory info (GL_ATI_meminfo):
    VBO free memory - total: 3648 MB, largest block: 3648 MB
    VBO free aux. memory - total: 3577 MB, largest block: 3577 MB
    Texture free memory - total: 3648 MB, largest block: 3648 MB
    Texture free aux. memory - total: 3577 MB, largest block: 3577 MB
    Renderbuffer free memory - total: 3648 MB, largest block: 3648 MB
    Renderbuffer free aux. memory - total: 3577 MB, largest block: 3577 MB
Memory info (GL_NVX_gpu_memory_info):
    Dedicated video memory: 4096 MB
    Total available memory: 8192 MB
    Currently available dedicated video memory: 3648 MB
OpenGL vendor string: X.Org
OpenGL renderer string: Radeon RX550/550 Series (POLARIS12, DRM 3.35.0, 5.4.0-4-amd64, LLVM 9.0.1)
OpenGL core profile version string: 4.5 (Core Profile) Mesa 19.3.3
OpenGL core profile shading language version string: 4.50
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 4.5 (Compatibility Profile) Mesa 19.3.3
OpenGL shading language version string: 4.50
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile

OpenGL ES profile version string: OpenGL ES 3.2 Mesa 19.3.3
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

le pilote openCL :

clinfo
Number of platforms                               1
  Platform Name                                   AMD Accelerated Parallel Processing
  Platform Vendor                                 Advanced Micro Devices, Inc.
  Platform Version                                OpenCL 2.1 AMD-APP (3004.6)
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd cl_amd_event_callback cl_amd_offline_devices 
  Platform Host timer resolution                  1ns
  Platform Extensions function suffix             AMD

  Platform Name                                   AMD Accelerated Parallel Processing
Number of devices                                 1
  Device Name                                     gfx804
  Device Vendor                                   Advanced Micro Devices, Inc.
  Device Vendor ID                                0x1002
  Device Version                                  OpenCL 1.2 AMD-APP (3004.6)
  Driver Version                                  3004.6
  Device OpenCL C Version                         OpenCL C 1.2 
  Device Type                                     GPU
  Device Board Name (AMD)                         Radeon RX550/550 Series
  Device Topology (AMD)                           PCI-E, 0a:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               8
  SIMD per compute unit (AMD)                     4
  SIMD width (AMD)                                16
  SIMD instruction width (AMD)                    1
  Max clock frequency                             1206MHz
  Graphics IP (AMD)                               8.0
  Device Partition                                (core)
    Max number of sub-devices                     8
    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                                     No
    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                              3816013824 (3.554GiB)
  Global free memory (AMD)                        3706992 (3.535GiB)
  Global memory channels (AMD)                    4
  Global memory banks per channel (AMD)           16
  Global memory bank width (AMD)                  256 bytes
  Error Correction support                        No
  Max memory allocation                           3054000332 (2.844GiB)
  Unified memory for Host and Device              No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       2048 bits (256 bytes)
  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             16
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 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                             2048x2048x2048 pixels
    Max number of read image args                 128
    Max number of write image args                8
  Local memory type                               Local
  Local memory size                               32768 (32KiB)
  Local memory syze per CU (AMD)                  65536 (64KiB)
  Local memory banks (AMD)                        32
  Max number of constant args                     8
  Max constant buffer size                        3054000332 (2.844GiB)
  Preferred constant buffer size (AMD)            16384 (16KiB)
  Max size of kernel argument                     1024
  Queue properties                                
    Out-of-order execution                        No
    Profiling                                     Yes
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      1ns
  Profiling timer offset since Epoch (AMD)        1586590360034232392ns (Sat Apr 11 09:32:40 2020)
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Thread trace supported (AMD)                  Yes
    Number of async queues (AMD)                  2
    Max real-time compute queues (AMD)            0
    Max real-time compute units (AMD)             0
    SPIR versions                                 1.2
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_fp64 cl_amd_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_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_image2d_from_buffer cl_khr_spir cl_khr_gl_event 

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                                   gfx804
  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                                   gfx804
  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                                   gfx804

AMD RX 570 8GB DDR5

ma nouvelle carte graphique :

lspci -nnkd::0300
0a:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] [1002:67df] (rev ef)
	Subsystem: Sapphire Technology Limited Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] [1da2:e343]
	Kernel driver in use: amdgpu
	Kernel modules: amdgpu

son pilote classique

glxinfo -B
name of display: :0
display: :0  screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
    Vendor: X.Org (0x1002)
    Device: Radeon RX 570 Series (POLARIS10, DRM 3.36.0, 5.5.0-1-amd64, LLVM 9.0.1) (0x67df)
    Version: 19.3.3
    Accelerated: yes
    Video memory: 8192MB
    Unified memory: no
    Preferred profile: core (0x1)
    Max core profile version: 4.5
    Max compat profile version: 4.5
    Max GLES1 profile version: 1.1
    Max GLES[23] profile version: 3.2
Memory info (GL_ATI_meminfo):
    VBO free memory - total: 7650 MB, largest block: 7650 MB
    VBO free aux. memory - total: 8140 MB, largest block: 8140 MB
    Texture free memory - total: 7650 MB, largest block: 7650 MB
    Texture free aux. memory - total: 8140 MB, largest block: 8140 MB
    Renderbuffer free memory - total: 7650 MB, largest block: 7650 MB
    Renderbuffer free aux. memory - total: 8140 MB, largest block: 8140 MB
Memory info (GL_NVX_gpu_memory_info):
    Dedicated video memory: 8192 MB
    Total available memory: 16384 MB
    Currently available dedicated video memory: 7650 MB
OpenGL vendor string: X.Org
OpenGL renderer string: Radeon RX 570 Series (POLARIS10, DRM 3.36.0, 5.5.0-1-amd64, LLVM 9.0.1)
OpenGL core profile version string: 4.5 (Core Profile) Mesa 19.3.3
OpenGL core profile shading language version string: 4.50
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 4.5 (Compatibility Profile) Mesa 19.3.3
OpenGL shading language version string: 4.50
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile

OpenGL ES profile version string: OpenGL ES 3.2 Mesa 19.3.3
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

pilote open CL

clinfo
Number of platforms                               1
  Platform Name                                   AMD Accelerated Parallel Processing
  Platform Vendor                                 Advanced Micro Devices, Inc.
  Platform Version                                OpenCL 2.1 AMD-APP (3004.6)
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd cl_amd_event_callback cl_amd_offline_devices 
  Platform Host timer resolution                  1ns
  Platform Extensions function suffix             AMD

  Platform Name                                   AMD Accelerated Parallel Processing
Number of devices                                 1
  Device Name                                     Ellesmere
  Device Vendor                                   Advanced Micro Devices, Inc.
  Device Vendor ID                                0x1002
  Device Version                                  OpenCL 1.2 AMD-APP (3004.6)
  Driver Version                                  3004.6
  Device OpenCL C Version                         OpenCL C 1.2 
  Device Type                                     GPU
  Device Board Name (AMD)                         Radeon RX 570 Series
  Device Topology (AMD)                           PCI-E, 0a:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               32
  SIMD per compute unit (AMD)                     4
  SIMD width (AMD)                                16
  SIMD instruction width (AMD)                    1
  Max clock frequency                             1244MHz
  Graphics IP (AMD)                               8.0
  Device Partition                                (core)
    Max number of sub-devices                     32
    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                                     No
    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                              8016703488 (7.466GiB)
  Global free memory (AMD)                        7809228 (7.447GiB)
  Global memory channels (AMD)                    8
  Global memory banks per channel (AMD)           16
  Global memory bank width (AMD)                  256 bytes
  Error Correction support                        No
  Max memory allocation                           4244635648 (3.953GiB)
  Unified memory for Host and Device              No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       2048 bits (256 bytes)
  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             16
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 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                             2048x2048x2048 pixels
    Max number of read image args                 128
    Max number of write image args                8
  Local memory type                               Local
  Local memory size                               32768 (32KiB)
  Local memory syze per CU (AMD)                  65536 (64KiB)
  Local memory banks (AMD)                        32
  Max number of constant args                     8
  Max constant buffer size                        4244635648 (3.953GiB)
  Preferred constant buffer size (AMD)            16384 (16KiB)
  Max size of kernel argument                     1024
  Queue properties                                
    Out-of-order execution                        No
    Profiling                                     Yes
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      1ns
  Profiling timer offset since Epoch (AMD)        1587542527048143648ns (Wed Apr 22 10:02:07 2020)
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Thread trace supported (AMD)                  Yes
    Number of async queues (AMD)                  2
    Max real-time compute queues (AMD)            0
    Max real-time compute units (AMD)             575513232
    SPIR versions                                 1.2
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_fp64 cl_amd_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_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_image2d_from_buffer cl_khr_spir cl_khr_gl_event 

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                                   Ellesmere
  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                                   Ellesmere
  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                                   Ellesmere

pilote 20.10 CL Info

clinfo
Number of platforms                               1
  Platform Name                                   AMD Accelerated Parallel Processing
  Platform Vendor                                 Advanced Micro Devices, Inc.
  Platform Version                                OpenCL 2.1 AMD-APP (3075.10)
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd cl_amd_event_callback cl_amd_offline_devices 
  Platform Host timer resolution                  1ns
  Platform Extensions function suffix             AMD

  Platform Name                                   AMD Accelerated Parallel Processing
Number of devices                                 1
  Device Name                                     Ellesmere
  Device Vendor                                   Advanced Micro Devices, Inc.
  Device Vendor ID                                0x1002
  Device Version                                  OpenCL 1.2 AMD-APP (3075.10)
  Driver Version                                  3075.10
  Device OpenCL C Version                         OpenCL C 1.2 
  Device Type                                     GPU
  Device Board Name (AMD)                         Radeon RX 570 Series
  Device Topology (AMD)                           PCI-E, 0b:00.0
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               32
  SIMD per compute unit (AMD)                     4
  SIMD width (AMD)                                16
  SIMD instruction width (AMD)                    1
  Max clock frequency                             1244MHz
  Graphics IP (AMD)                               8.0
  Device Partition                                (core)
    Max number of sub-devices                     32
    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                                     No
    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                              8022630400 (7.472GiB)
  Global free memory (AMD)                        7815016 (7.453GiB)
  Global memory channels (AMD)                    8
  Global memory banks per channel (AMD)           16
  Global memory bank width (AMD)                  256 bytes
  Error Correction support                        No
  Max memory allocation                           4244635648 (3.953GiB)
  Unified memory for Host and Device              No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       2048 bits (256 bytes)
  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             16
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 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                             2048x2048x2048 pixels
    Max number of read image args                 128
    Max number of write image args                8
  Local memory type                               Local
  Local memory size                               32768 (32KiB)
  Local memory syze per CU (AMD)                  65536 (64KiB)
  Local memory banks (AMD)                        32
  Max number of constant args                     8
  Max constant buffer size                        4244635648 (3.953GiB)
  Preferred constant buffer size (AMD)            16384 (16KiB)
  Max size of kernel argument                     1024
  Queue properties                                
    Out-of-order execution                        No
    Profiling                                     Yes
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      1ns
  Profiling timer offset since Epoch (AMD)        1591853725436918849ns (Thu Jun 11 07:35:25 2020)
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Thread trace supported (AMD)                  Yes
    Number of async queues (AMD)                  2
    Max real-time compute queues (AMD)            0
    Max real-time compute units (AMD)             909198854
    SPIR versions                                 1.2
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_fp64 cl_amd_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_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_image2d_from_buffer cl_khr_spir cl_khr_gl_event 

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                                   Ellesmere
  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                                   Ellesmere
  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                                   Ellesmere

Solution Alternative

testé sur debian testing bullseye (kernel 5.9.0-1) le paquet et ses dépendances s'installe très bien mais n'est pas fonctionnel dans blender 2.83.5

apt search mesa-opencl-icd
En train de trier... Fait
Recherche en texte intégral... Fait
mesa-opencl-icd/testing,now 20.1.9-1 amd64  [installé]
  free implementation of the OpenCL API -- ICD runtime
apt show mesa-opencl-icd
Package: mesa-opencl-icd
Version: 20.1.9-1
Priority: optional
Section: libs
Source: mesa
Maintainer: Debian X Strike Force <debian-x@lists.debian.org>
Installed-Size: 29,4 MB
Provides: opencl-icd
Depends: libclc-r600 (>= 0.2.0+git20180312-1~), libclc-amdgcn (>= 0.2.0+git20180312-1~), ocl-icd-libopencl1 | libopencl1, libc6 (>= 2.29), libclang-cpp10, libdrm-amdgpu1 (>= 2.4.100), libdrm-nouveau2 (>= 2.4.66), libdrm-radeon1 (>= 2.4.31), libdrm2 (>= 2.4.75), libelf1 (>= 0.142), libexpat1 (>= 2.0.1), libgcc-s1 (>= 3.4), libllvm10 (>= 1:9~svn298832-1~), libstdc++6 (>= 5.2), libzstd1 (>= 1.3.2), zlib1g (>= 1:1.1.4)
Homepage: https://mesa3d.org/
Tag: role::shared-lib
Download-Size: 4809 kB
APT-Manual-Installed: yes
APT-Sources: http://deb.debian.org/debian testing/main amd64 Packages
Description: free implementation of the OpenCL API -- ICD runtime
 This package contains the mesa implementation of the OpenCL (Open Compute
 Language) library, which is intended for use with an ICD loader. OpenCL
 provides a standardized interface for computational analysis on graphical
 processing units.
1)
N'hésitez pas à y faire part de vos remarques, succès, améliorations ou échecs !
2)
En informatique, le calcul parallèle consiste en l'exécution simultanée d'une même tâche, partitionnée et adaptée afin de pouvoir être répartie entre plusieurs processeurs en vue de traiter plus rapidement des problèmes plus grands.
3)
stable
4)
sous testing bullseye kernel 5.8.0-2 , par exemple
5)
Copie.sh
6)
ou à la souris
7)
utilisateur
8)
actuellement réservés à root
atelier/chantier/opencl-avec-amd.1630047810.txt.gz · Dernière modification: 27/08/2021 09:03 par Debian Alain

Pied de page des forums

Propulsé par FluxBB