[Beignet] troubles with beignet 0.8.0, Debian 8 (jessie) and i7 4770

Zhigang Gong zhigang.gong at linux.intel.com
Fri Jun 13 02:50:31 PDT 2014


About the ICD support, you can read the README.md, I just paste it as below
for your reference:

  And if you want to work with the standard ICD libOpenCL.so, then you need
  two more packages (the following package name is for Ubuntu):

  - ocl-icd-dev
  - ocl-icd-libopencl1

  If you don't want to enable ICD, or your system doesn't have ICD OpenCL
support,
  you can still link to the beignet OpenCL library. You can find the
beignet/libcl.so
  in your system's library installation directories.

Please be noted. After you install the above two packages, you need to do a
complete clean
build on beignet from a clean directory

cmake ../beignet
make
sudo make install

After that you need to check whether the intel_beignet is installed at (for
Ubuntu system):
/etc/OpenCL/vendors/intel-beignet.icd


> -----Original Message-----
> From: Beignet [mailto:beignet-bounces at lists.freedesktop.org] On Behalf Of
> Nikita Orlov
> Sent: Friday, June 13, 2014 5:18 PM
> To: Yang, Rong R; Yichao Yu
> Cc: beignet at lists.freedesktop.org
> Subject: Re: [Beignet] troubles with beignet 0.8.0, Debian 8 (jessie) and
i7 4770
> 
> By the way!
> 
> Now clinfo gives next info:
> I: ICD loader reports no usable platforms.
> 
> I think I really don`t complete some configs :-(
> 
> 13.06.2014, 13:11, "Nikita Orlov" <nikitos1550 at yandex.ru>:
> > Hi!
> >
> > Thanks very much!
> > I have built beignet from master branch and patch kernel 3.14.
> > Now beignet utest pass 0.199 (5 tests failed).
> >
> > But, I still have a problem, OpenCV still fail all tests from
opencv_test_ocl.
> > I think I must config something, but I can`t find any doc, articles of
anything
> else about it.
> > Maybe you can point me right source for such information?
> >
> > About, HSW support. How do you think when will it be in upstream?
> >
> > 13.06.2014, 06:41, "Yang, Rong R" <rong.r.yang at intel.com>:
> >>  -----Original Message-----
> >>  From: Beignet [mailto:beignet-bounces at lists.freedesktop.org] On Behalf
> Of Yichao Yu
> >>  Sent: Friday, June 13, 2014 2:09 AM
> >>  To: Nikita Orlov
> >>  Cc: beignet at lists.freedesktop.org
> >>  Subject: Re: [Beignet] troubles with beignet 0.8.0, Debian 8 (jessie)
and i7
> 4770
> >>
> >>  On Thu, Jun 12, 2014 at 12:52 PM, Nikita Orlov <nikitos1550 at yandex.ru>
> wrote:
> >>>   Hello!
> >>>
> >>>   I`m trying to run opencv opencl accelerated.
> >>>   I have installed beignet from debian jessie repo.
> >>>
> >>>   clinfo:
> >>>
> >>>   user at video-stend2:~$ clinfo
> >>>   Number of platforms: 1
> >>>     Platform Profile: FULL_PROFILE
> >>>     Platform Version: OpenCL 1.1 beignet 0.8.0
> >>>     Platform Name: Experiment 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_icd
> >>>
> >>>     Platform Name: Experiment Intel Gen OCL Driver Number of devices:
> 1
> >>>     Device Type: CL_DEVICE_TYPE_GPU
> >>>     Device ID: 1042
> >>>     Max compute units: 64
> >>>     Max work items dimensions: 3
> >>>       Max work items[0]: 512
> >>>       Max work items[1]: 512
> >>>       Max work items[2]: 512
> >>>     Max work group size: 512
> >>>     Preferred vector width char: 16
> >>>     Preferred vector width short: 16
> >>>     Preferred vector width int: 16
> >>>     Preferred vector width long: 16
> >>>     Preferred vector width float: 16
> >>>     Preferred vector width double: 0
> >>>     Native vector width char: 16
> >>>     Native vector width short: 16
> >>>     Native vector width int: 16
> >>>     Native vector width long: 16
> >>>     Native vector width float: 16
> >>>     Native vector width double: 16
> >>>     Max clock frequency: 1000Mhz
> >>>     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: 8192
> >>>     Max image 2D height: 8192
> >>>     Max image 3D width: 8192
> >>>     Max image 3D height: 8192
> >>>     Max image 3D depth: 2048
> >>>     Max samplers within kernel: 16
> >>>     Max size of kernel argument: 1024
> >>>     Alignment (bits) of base address: 1024
> >>>     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: No
> >>>       Round to +ve and infinity: No
> >>>       IEEE754-2008 fused multiply-add: No
> >>>     Cache type: Read/Write
> >>>     Cache line size: 128
> >>>     Cache size: 8192
> >>>     Global memory size: 134217728
> >>>     Constant buffer size: 524288
> >>>     Max number of constant args: 8
> >>>     Local memory type: Global
> >>>     Local memory size: 65536
> >>>     Error correction support: 0
> >>>     Unified memory for Host and Device: 0
> >>>     Profiling timer resolution: 80
> >>>     Device endianess: Little
> >>>     Available: Yes
> >>>     Compiler available: Yes
> >>>     Execution capabilities:
> >>>       Execute OpenCL kernels: Yes
> >>>       Execute native function: Yes
> >>>     Queue properties:
> >>>       Out-of-Order: No
> >>>       Profiling : Yes
> >>>     Platform ID: 0x7f1390957a40
> >>>     Name: Intel(R) HD Graphics Haswell M
> >>>     Vendor: Intel
> >>>     Device OpenCL C version: OpenCL C 1.1 beignet 0.8.0
> >>>     Driver version: 0.8.0
> >>>     Profile: FULL_PROFILE
> >>>     Version: OpenCL 1.1 beignet 0.8.0
> >>>     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_icd
> >>>
> >>>   user at video-stend2:~$ clinfo
> >>>   Number of platforms: 1
> >>>     Platform Profile: FULL_PROFILE
> >>>     Platform Version: OpenCL 1.1 beignet 0.8.0
> >>>     Platform Name: Experiment 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_icd
> >>>
> >>>     Platform Name: Experiment Intel Gen OCL Driver Number of devices:
> 1
> >>>     Device Type: CL_DEVICE_TYPE_GPU
> >>>     Device ID: 1042
> >>>     Max compute units: 64
> >>>     Max work items dimensions: 3
> >>>       Max work items[0]: 512
> >>>       Max work items[1]: 512
> >>>       Max work items[2]: 512
> >>>     Max work group size: 512
> >>>     Preferred vector width char: 16
> >>>     Preferred vector width short: 16
> >>>     Preferred vector width int: 16
> >>>     Preferred vector width long: 16
> >>>     Preferred vector width float: 16
> >>>     Preferred vector width double: 0
> >>>     Native vector width char: 16
> >>>     Native vector width short: 16
> >>>     Native vector width int: 16
> >>>     Native vector width long: 16
> >>>     Native vector width float: 16
> >>>     Native vector width double: 16
> >>>     Max clock frequency: 1000Mhz
> >>>     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: 8192
> >>>     Max image 2D height: 8192
> >>>     Max image 3D width: 8192
> >>>     Max image 3D height: 8192
> >>>     Max image 3D depth: 2048
> >>>     Max samplers within kernel: 16
> >>>     Max size of kernel argument: 1024
> >>>     Alignment (bits) of base address: 1024
> >>>     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: No
> >>>       Round to +ve and infinity: No
> >>>       IEEE754-2008 fused multiply-add: No
> >>>     Cache type: Read/Write
> >>>     Cache line size: 128
> >>>     Cache size: 8192
> >>>     Global memory size: 134217728
> >>>     Constant buffer size: 524288
> >>>     Max number of constant args: 8
> >>>     Local memory type: Global
> >>>     Local memory size: 65536
> >>>     Error correction support: 0
> >>>     Unified memory for Host and Device: 0
> >>>     Profiling timer resolution: 80
> >>>     Device endianess: Little
> >>>     Available: Yes
> >>>     Compiler available: Yes
> >>>     Execution capabilities:
> >>>       Execute OpenCL kernels: Yes
> >>>       Execute native function: Yes
> >>>     Queue properties:
> >>>       Out-of-Order: No
> >>>       Profiling : Yes
> >>>     Platform ID: 0x7f1390957a40
> >>>     Name: Intel(R) HD Graphics Haswell M
> >>>     Vendor: Intel
> >>>     Device OpenCL C version: OpenCL C 1.1 beignet 0.8.0
> >>>     Driver version: 0.8.0
> >>>     Profile: FULL_PROFILE
> >>>     Version: OpenCL 1.1 beignet 0.8.0
> >>>     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_icd ccc
> >>>
> >>>   I am trying run opencv_test_ocl and utest (from beignet src, I can
> succesfully build ut by myself). Everything fails.
> >>>   It is difficult to paste here log, because system freeze and I can
just
> reset computer.
> >>  It seems that you have a Haswell CPU and I think you are having the
same
> problem with me (You can find a thread on the mailing list a few weeks ago
> about it). The support for Haswell CPU is not finished yet and it will
very likely
> hang the GPU.
> >>>>>>>   Are you using beignet 0.8 release? Beignet v0.8 have not
> supported HSW, you should use Beignet git master, the Haswell support
almost
> has finished in master. Pay attention that you should apply a kernel patch
in
> attachment if you want to run kernels with local memory. This is because
of
> drm's restriction, and will drm guys is handing it.
> >>  If still any problem, welcome to report.
> >>
> >>  P.S. I don't think the system is frozen. It might appear that way
because
> the display is not updating. However, if you wait long enough and/or try
to kill
> the running process with Ctrl-C, you can still get you system back.
> >>>>>>>>  System frozen may be caused by GPU hang and your system
> doesn't enable hang check. Try echo -n 1 >
> /sys/module/i915/parameters/enable_hangcheck with root user.
> >>>   I run tests with Xorg and without.
> >>>
> >>>   What is wrong? Maybe I didn`t install some libs or some additional
> configuration is required?
> >>>
> >>>   --
> >>>   Nikita Orlov
> >>>   Skype: nik_stet
> >>>   QQ: 2717846083
> >>>   _______________________________________________
> >>>   Beignet mailing list
> >>>   Beignet at lists.freedesktop.org
> >>>   http://lists.freedesktop.org/mailman/listinfo/beignet
> >>  _______________________________________________
> >>  Beignet mailing list
> >>  Beignet at lists.freedesktop.org
> >>  http://lists.freedesktop.org/mailman/listinfo/beignet
> >>  ,
> >>
> >>  _______________________________________________
> >>  Beignet mailing list
> >>  Beignet at lists.freedesktop.org
> >>  http://lists.freedesktop.org/mailman/listinfo/beignet
> >
> > --
> > Nikita Orlov
> > Skype: nik_stet
> > QQ: 2717846083
> > _______________________________________________
> > Beignet mailing list
> > Beignet at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/beignet
> 
> --
> Nikita Orlov
> Skype: nik_stet
> QQ: 2717846083
> _______________________________________________
> Beignet mailing list
> Beignet at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/beignet



More information about the Beignet mailing list