helloworld_ocl fails with Error(-45)

[Old posts from the commercial version of ArrayFire] Issues and comments for download and installation. Getting up and running.

Moderators: melonakos, pavanky

helloworld_ocl fails with Error(-45)

Postby a-loc » Wed Nov 20, 2013 7:28 am

Hello,

I have a NVIDIA Quadro NVS 160M. I know that it's a pretty old GPU. helloworld_cuda works though. What could be the problem?

Code: Select all
$ ./helloworld_ocl
ArrayFire v2.0.RC1  (OpenCL, 64-bit Linux, build b1809d9)
License: Standalone (/opt/arrayfire/2.0/arrayfire.lic)
Platform: NVIDIA CUDA, Driver: 319.37
[0]: Quadro NVS 160M, 255 MB, OpenCL Version: 1.0

create a 5-by-3 matrix of random floats on the GPU
src/opencl/rand.cpp:39: _err  -45
src/opencl/rand.cpp:39: src/opencl/rand.cpp:39 Error(-45)
terminate called after throwing an instance of 'af::exception'
Aborted


Code: Select all
$ ./helloworld_cuda
ArrayFire v2.0.RC1  (CUDA, 64-bit Linux, build b1809d9)
License: Standalone (/opt/arrayfire/2.0/arrayfire.lic)
Platform: CUDA toolkit 5.5, Driver: 319.37
 0 : Quadro NVS 160M, 256 MB, CUDA Compute 1.1 (single)
Memory Usage: 99 MB free (256 MB total)

create a 5-by-3 matrix of random floats on the GPU
A =
        0.7402     0.4464     0.7762
        0.9210     0.6673     0.2948
        0.0390     0.1099     0.7140
        0.9690     0.4702     0.3585
        0.9251     0.5132     0.6814

element-wise arithmetic
B =
        0.7744     0.5317     0.8006
        0.8962     0.7189     0.3905
        0.1390     0.2097     0.7549
        0.9243     0.5531     0.4509
        0.8987     0.5910     0.7299

Fourier transform the result
C =
           3.6327 + 0.0000i       2.6043 + 0.0000i       3.1267 + 0.0000i
           0.4689 + 0.4640i       0.3193 + 0.0802i       0.1713 + 0.1441i
          -0.3491 - 0.7454i      -0.2923 - 0.4018i       0.2667 + 0.4886i
          -0.3491 + 0.7454i      -0.2923 + 0.4018i       0.2667 - 0.4886i
           0.4689 - 0.4640i       0.3193 - 0.0802i       0.1713 - 0.1441i

grab last row
c =
           0.4689 - 0.4640i       0.3193 - 0.0802i       0.1713 - 0.1441i

zero out every other column
negate the first three elements of middle column
B =
        0.0000    -0.5317     0.0000
        0.0000    -0.7189     0.0000
        0.0000    -0.2097     0.0000
        0.0000     0.5531     0.0000
        0.0000     0.5910     0.0000

create 2-by-3 matrix from host data
D =
        1.0000     3.0000     5.0000
        2.0000     4.0000     6.0000

copy last column onto first
D =
        5.0000     3.0000     5.0000
        6.0000     4.0000     6.0000


Code: Select all
$ ./clinfo
Number of platforms:             1
  Platform Profile:             FULL_PROFILE
  Platform Version:             OpenCL 1.1 CUDA 4.2.1
  Platform Name:             NVIDIA CUDA
  Platform Vendor:             NVIDIA Corporation
  Platform Extensions:             cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll

  Platform Name:             NVIDIA CUDA
Number of devices:             1
  Device Type:                CL_DEVICE_TYPE_GPU
  Device ID:                4318
  Max compute units:             1
  Max work items dimensions:          3
    Max work items[0]:             512
    Max work items[1]:             512
    Max work items[2]:             64
  Max work group size:             512
  Preferred vector width char:          1
  Preferred vector width short:          1
  Preferred vector width int:          1
  Preferred vector width long:          1
  Preferred vector width float:          1
  Preferred vector width double:       0
  Native vector width char:          1
  Native vector width short:          1
  Native vector width int:          1
  Native vector width long:          1
  Native vector width float:          1
  Native vector width double:          0
  Max clock frequency:             1450Mhz
  Address bits:                32
  Max memory allocation:          134217728
  Image support:             Yes
  Max number of images read arguments:       128
  Max number of images write arguments:       8
  Max image 2D width:             4096
  Max image 2D height:             16383
  Max image 3D width:             2048
  Max image 3D height:             2048
  Max image 3D depth:             2048
  Max samplers within kernel:          16
  Max size of kernel argument:          4352
  Alignment (bits) of base address:       2048
  Minimum alignment (bytes) for any datatype:    128
  Single precision floating point capability
    Denorms:                No
    Quiet NaNs:                Yes
    Round to nearest even:          Yes
    Round to zero:             Yes
    Round to +ve and infinity:          Yes
    IEEE754-2008 fused multiply-add:       Yes
  Cache type:                None
  Cache line size:             0
  Cache size:                0
  Global memory size:             268107776
  Constant buffer size:             65536
  Max number of constant args:          9
  Local memory type:             Scratchpad
  Local memory size:             16384
  Kernel Preferred work group size multiple:    32
  Error correction support:          0
  Unified memory for Host and Device:       0
  Profiling timer resolution:          1000
  Device endianess:             Little
  Available:                Yes
  Compiler available:             Yes
  Execution capabilities:            
    Execute OpenCL kernels:          Yes
    Execute native function:          No
  Queue properties:            
    Out-of-Order:             Yes
    Profiling :                Yes
  Platform ID:                0x00000000011bdfc0
  Name:                   Quadro NVS 160M
  Vendor:                NVIDIA Corporation
  Device OpenCL C version:          OpenCL C 1.0
  Driver version:             319.37
  Profile:                FULL_PROFILE
  Version:                OpenCL 1.0 CUDA
  Extensions:                cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll  cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics
a-loc
 
Posts: 3
Joined: Wed Nov 20, 2013 7:03 am

Re: helloworld_ocl fails with Error(-45)

Postby pavanky » Tue Dec 03, 2013 5:26 pm

Hi,

Can you change randu(5,3) to constant(0, 5, 3) and see if anything else breaks ?

We are testing on an OpenCL 1.0 card (from NVIDIA) and the same code does not seem to break. We'll try to find another GPU which is closer to the one you are using and see how far we will get.
Pavan Yalamanchili,
ArrayFire
--
~ If it is not broken, you have not tried hard enough ~
User avatar
pavanky
Site Admin
 
Posts: 1123
Joined: Mon Mar 15, 2010 7:39 pm
Location: Atlanta, GA


Return to [archive-commercial] Download & Installation

cron