CUDA OpenCL and openmp

Front page Forums Installation CUDA OpenCL and openmp

This topic contains 4 replies, has 2 voices, and was last updated by  nico 9 months, 4 weeks ago.

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #1469

    Najib
    Member

    Hi,
    I have some problems with paralution-1.1.0:
    1st parallel computing :
    ——————————–
    i installed the library with the default cmake command (cmake ..) no CUDA neither OpenCL, it work but i get always the same computing time when i modify the number of threads (set_omp_threads_paralution(n) (No MKL)
    2nd CUDA :
    ————–
    cuda is installed (ubuntu 16.04 i don’t know if there is a problem of compatibity with my gpu this is the output message identifiing gpu :
    …………………………………………………………………………………
    description: VGA compatible controller
    product: Juniper XT [FirePro V5800]
    vendor: Advanced Micro Devices, Inc. [AMD/ATI]
    ………………………………………………………………………………….
    )

    and then i installed the library with (cmake -DSUPPORT_CUDA=ON ..)
    i got this message

    Number of CPU cores: 12
    Host thread affinity policy – thread mapping on every core
    Querying for GPU devices failed – falling back to host backend
    Number of CPU cores: 12
    Host thread affinity policy – thread mapping on every core
    PARALUTION ver B1.1.0
    PARALUTION platform is initialized
    Accelerator backend: GPU(CUDA)
    OpenMP threads:12
    GPU is not initialized
    ————————————

    it is similar with OpenCL :
    ———————————-
    Number of CPU cores: 12
    Host thread affinity policy – thread mapping on every core
    No OpenCL platform available
    OpenCL has NOT been initialized!
    Number of CPU cores: 12
    Host thread affinity policy – thread mapping on every core
    PARALUTION ver B1.1.0
    PARALUTION platform is initialized
    Accelerator backend: OpenCL
    OpenMP threads:12
    OpenCL is not initialized
    ————————————–

    any helps please
    best regards
    Najib

    #1470

    nico
    Participant

    Hi Najib,

    Regarding the CPU scalability: Is your test case sufficiently large? What kind of CPU are you using, is it possibly, that your memory bandwidth is already fully utilized by very few cores?
    You are trying to run the CUDA backend on AMD hardware, this is not possible since CUDA is only supported by NVIDIA hardware.
    Regarding OpenCL, can you make sure that the OpenCL runtime environment is properly configured and working? You can test this by running some opencl samples supplied with the AMD opencl sdk or by checking clinfo.

    Best
    Nico

    #1471

    Najib
    Member

    Thank s Nico
    About memory i have sufficient quota ( 50go free) i made à square matrix of size 1000 for testing !
    I Will look for opencl
    Thank you again

    #1472

    Najib
    Member

    Hi,

    About OpencL : this output message of clinfo:
    ——————————————————————————————–
    DRM_IOCTL_I915_GEM_APERTURE failed: No such file or directory
    Assuming 131072kB available aperture size.
    May lead to reduced performance or incorrect rendering.
    get chip id failed: -1 [22]
    param: 4, val: 0
    Number of platforms: 1
    Platform Profile: FULL_PROFILE
    Platform Version: OpenCL 1.2 beignet 1.3
    Platform Name: Intel Gen OCL Driver
    Platform Vendor: Intel
    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_byte_addressable_store cl_khr_3d_image_writes cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_spir cl_khr_icd cl_intel_accelerator cl_intel_subgroups cl_intel_subgroups_short

    Platform Name: Intel Gen OCL Driver
    DRM_IOCTL_I915_GEM_APERTURE failed: No such file or directory
    Assuming 131072kB available aperture size.
    May lead to reduced performance or incorrect rendering.
    get chip id failed: -1 [22]
    param: 4, val: 0
    cl_get_gt_device(): error, unknown device: 0
    ERROR: clGetDeviceIDs(-1)
    —————————————————————————————————————————————–
    for openmp : “What kind of CPU are you using ?”
    i have Intel(R) Xeon(R) CPU X5675 @ 3.07GHz
    is that needs a intel c++ compiler ? consequently i have to use SUPPORT_MIC ?

    best

    #1473

    nico
    Participant

    Hi Najib,

    looks like there is something wrong with your OpenCL SDK installation. Please make sure that everything related to OpenCL works, before moving on to PARALUTION.

    Regarding the OpenMP test you did, please try again with a larger case. You can check the UF Sparse Matrix Collection for example to find a larger test case.

    Best
    Nico

Viewing 5 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic.