AnsweredAssumed Answered

APU's GPU cannot be detected for OpenCL

Question asked by lennox on Sep 7, 2015

Hi guys,

I want to use APU's(A10-7870K) SVM ability.

I installed the latest driver(15.20.1046) and the SDK3.0, however, the clinfo show the supported Opencl version is 1.2 and the SVM is not supported.

Actuallly, the GPU is not detected, right?

I'm sure the spec of 7870K show the APU can support OPENCL2.0, right?

The OS which I use is OpenSuse 13.2, and I also try the ubuntu 15, 14, and Windows 10.

Windows 10 can detect the GPU and suppoort SVM, so the setting of BIOS should be OK.

I have tried AMD OpenCL™ 2.0 Driver , however, the version seem too old for Opensuse 13.2.







Append the output of clinfo


  Platform Name:                            AMD Accelerated Parallel Processing
Number of devices:                          1
  Device Type:                              CL_DEVICE_TYPE_CPU
  Vendor ID:                                1002h
  Board name:                             
  Max compute units:                        4
  Max work items dimensions:                3
Max work items[0]:                      1024
Max work items[1]:                      1024
Max work items[2]:                      1024
  Max work group size:                      1024
  Preferred vector width char:              16
  Preferred vector width short:             8
  Preferred vector width int:               4
  Preferred vector width long:              2
  Preferred vector width float:             8
  Preferred vector width double:            4
  Native vector width char:                 16
  Native vector width short:                8
  Native vector width int:                  4
  Native vector width long:                 2
  Native vector width float:                8
  Native vector width double:               4
  Max clock frequency:                      1700Mhz
  Address bits:                             64
  Max memory allocation:                    8154801152
  Image support:                            Yes
  Max number of images read arguments:      128
  Max number of images write arguments:     64
  Max image 2D width:                       8192
  Max image 2D height:                      8192
  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:              4096
  Alignment (bits) of base address:         1024
  Minimum alignment (bytes) for any datatype:128

  Single precision floating point capability

Denorms:                                Yes
Quiet NaNs:                             Yes
Round to nearest even:                  Yes
Round to zero:                          Yes
Round to +ve and infinity:              Yes
IEEE754-2008 fused multiply-add:        Yes
  Cache type:                               Read/Write
  Cache line size:                          64
  Cache size:                               16384
  Global memory size:                       32619204608
  Constant buffer size:                     65536
  Max number of constant args:              8
  Local memory type:                        Global
  Local memory size:                        32768
  Max pipe arguments:                       16
  Max pipe active reservations:             16
  Max pipe packet size:                     3859833856
  Max global variable size:                 1879048192
  Max global variable preferred total size: 1879048192
  Max read/write image args:                64
  Max on device events:                     0
  Queue on device max size:                 0
  Max on device queues:                     0
  Queue on device preferred size:           0
  SVM capabilities:                       
Coarse grain buffer:                    No
Fine grain buffer:                      No
Fine grain system:                      No
Atomics:                                No
  Preferred platform atomic alignment:      0
  Preferred global atomic alignment:        0
  Preferred local atomic alignment:         0
  Kernel Preferred work group size multiple:1
  Error correction support:                 0
  Unified memory for Host and Device:       1
  Profiling timer resolution:               1
  Device endianess:                         Little
  Available:                                Yes
  Compiler available:                       Yes
  Execution capabilities:                         
Execute OpenCL kernels:                 Yes
Execute native function:                Yes
  Queue on Host properties:                       
Out-of-Order:                           No
Profiling :                             Yes
  Queue on Device properties:                     
Out-of-Order:                           No
Profiling :                             No
  Platform ID:                              0x7f585bfaa430
  Name:                                     AMD A10-7870K Radeon R7, 12 Compute Cores 4C+8G
  Vendor:                                   AuthenticAMD
  Device OpenCL C version:                  OpenCL C 1.2
  Driver version:                           1800.5 (sse2,avx,fma4)
  Profile:                                  FULL_PROFILE
  Version:                                  OpenCL 1.2 AMD-APP (1800.5)
  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_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_spir cl_khr_gl_event