Pages

Tuesday, April 19, 2011

Install OpenCL and PyOpenCL in Scientific Linux 6 (RHEL 6, CEntOS 6)

My rig is 64 bit system with RedHat 6 based OS, but my fera not supported for OpenCL using VGA(3200), so it using SSE2 from L310 processor  :(

Note that my Scientific Linux 6 installed from a 64 bit LiveDVD version, so might be a different step if using LiveCD.

Installing AMD APP:
Test using a sample
$ ./HelloCL
HelloCL!
Getting Platform Information
Creating a context AMD platform
Getting device info
Loading and compiling CL source
Running CL program
Done
Passed!

Installing PyOpenCl:
  • download PyOpenCL from PyOpenCL
  • make sure C++ compiler(g++), python, numpy, C++ boost library are installed
  • Tricky when execute configure.py: python configure.py   --cl-inc-dir=$HOME/AMD-APP-SDK-v2.4-lnx64/include   --cl-lib-dir=$HOME/AMD-APP-SDK-v2.4-lnx64/lib/x86_64   --cl-libname=OpenCL
  • then just make -> su -c "make install"

Result Test Running examples benchmark-all.py
$ ./benchmark-all.py
('Execution time of test without OpenCL: ', 34.952448844909668, 's')
=============================================
('Platform name:', 'AMD Accelerated Parallel Processing')
('Platform profile:', 'FULL_PROFILE')
('Platform vendor:', 'Advanced Micro Devices, Inc.')
('Platform version:', 'OpenCL 1.1 AMD-APP-SDK-v2.4 (595.10)')
---------------------------------------------------------------
('Device name:', 'AMD Athlon(tm) X2 Dual Core Processor L310')
('Device type:', 'CPU')
('Device memory: ', 3706, 'MB')
('Device max clock speed:', 1200, 'MHz')
('Device compute units:', 2)
Execution time of test: 0.00733989 s
Results OK

7 comments:

  1. This was very helpful. Thanks!

    ReplyDelete
  2. @Unknown: Glad if this helpful... :)

    ReplyDelete
  3. Replies
    1. yes check for the newest version at AMD OpenCL site

      Delete
  4. Does the CodeXL works for you? I can't use it.

    ReplyDelete
  5. Link to the APP SDK is now broken can be found here now

    http://developer.amd.com/tools-and-sdks/heterogeneous-computing/amd-accelerated-parallel-processing-app-sdk/downloads/#one

    ReplyDelete