Forum Discussion

davidwigley's avatar
davidwigley
Icon for New Contributor rankNew Contributor
4 years ago
Solved

Failed to initialize user clock?

Hello,

I am following the instructions from here:

https://github.com/intel/FPGA-Devcloud/tree/master/main/QuickStartGuides/OpenCL_Program_PAC_Quickstart/Arria%2010

And the programming of the device is failing. Has anyone seen this?

The diagnostics appear fine.

u75982@s005-n007:~/A10_OPENCL_AFU/v1.2.1/hello_world/bin$ aocl program acl0 hello_world_fpga_unsigned.aocx
aocl program: Running program from /opt/a10/inteldevstack/a10_gx_pac_ias_1_2_1_pv/opencl/opencl_bsp/linux64/libexec
libopae-c user_clk_pgm_uclock.c:200:set_userclock() **ERROR** : Failed to initialize user clock
libopae-c reconf.c:244:set_afu_userclock() **ERROR** : Failed to set user clock
libopae-c reconf.c:393:fpgaReconfigureSlot() **ERROR** : Failed to set user clock
Error writing bitstream to FPGA: not supported
Error programming device
aocl program: Program failed.

  • Hi @davidwigley,

    Thank you for posting in Intel community forum and hope all is well.
    I have tried the instruction in the similar nodes mention and are able to program successfully. (i.e. complete steps 3.6)


    Hence suspecting there might be some missing steps, would suggest to to check on the following:
    1) Selecting the correct development stack source which is 1.2.1.
    2) Recompile the program by running make clean followed by make command.
    3) Are you able to compile the opencl code into the FPGA executable successfully?
    Please do try and let us know if that helps.

    Best Wishes
    BB

6 Replies

  • Hi @davidwigley,

    Thank you for posting in Intel community forum and hope all is well.
    I have tried the instruction in the similar nodes mention and are able to program successfully. (i.e. complete steps 3.6)


    Hence suspecting there might be some missing steps, would suggest to to check on the following:
    1) Selecting the correct development stack source which is 1.2.1.
    2) Recompile the program by running make clean followed by make command.
    3) Are you able to compile the opencl code into the FPGA executable successfully?
    Please do try and let us know if that helps.

    Best Wishes
    BB

    • davidwigley's avatar
      davidwigley
      Icon for New Contributor rankNew Contributor

      Thank you @BoonBengT_Altera.

      I just tried rerunning the executable I built again following the same steps and it has worked.

      Not sure what changed on the backend, but the issue appears to be fixed now.

      Thank you for your help!

  • Hi @davidwigley,

    Great, good to know that it is working now, with no further clarification on this thread, it will be transitioned to community support where community users will further help you with doubts in this thread.
    Thanks for your questions and as always pleasure having you here.

    Best Wishes
    BB

  • davidwigley's avatar
    davidwigley
    Icon for New Contributor rankNew Contributor

    Now the issue occurs after programming the device, without error.

    Upon execution, it terminates by saying:

    Binary length is 15988 for 1 devices
    Context callback: Malformed program binary: missing .acl.board section
    Context callback: kernel argument info is not available
    Context callback: Invalid binary
    ERROR: CL_INVALID_BINARY
    Location: ../common/src/AOCLUtils/opencl.cpp:408
    Failed to create program with binary

    Has anyone seen this before?

  • davidwigley's avatar
    davidwigley
    Icon for New Contributor rankNew Contributor

    Hi @BoonBengT_Altera ,

    Appreciate the response and help. I have reviewed that post and am still stuck. My issue is not with running the program in emulation mode, that works.

    Do you know what else causes the warning Context callback: Malformed program binary: missing .acl.board section to occur?

    Apologies if this is a silly question but I am not very familiar with OpenCL nor DevCloud.

    I do not believe it is from my development environment since I running out of the DevCloud, but I am not sure.

    The board diagnostics appear ok, programming yields no errors, and this issue is persistent across nodes.

    Best,

    David