I installed the Intel PAC recently and I followed the instructions to install the drivers. Now, when I wanted to initialize the bsp with hello_world.aocx I will get the following errors:
aocl program: Running program from /opt/inteldevstack/a10_gx_pac_ias_1_2_1_pv/opencl/opencl_bsp/linux64/libexec
libopae-c bitstream.c:387:validate_bitstream_metadata() **ERROR** : Interface ID check failed
Error writing bitstream to FPGA: invalid parameter
Error programming device
aocl program: Program failed.
Also, this is the PR interface id: 69528db6-eb31-577a-8c36-68f9faa081f6
I was wondering if anybody found the solution to this issue?
I am using Intel PAC A10, but not using opencl.
Can you run other design example in the location root/inteldevstack/a10_gx_pac_ias_1_2_pv/hw/samples/? like hello_ afu.
Also the best way is to describe you whole information incluidng OS ,LINUX version etc. and what tests you have done?
Thanks for your response Jim. I was able to fix the issue by updating the firmware. However, the new issue I faced is that, I can run the opencl hello world example on the fpga and it worked fine but after couple of minutes (after I started the machine), the machine cannot recognize the fpga anymore, unless I power off the system and start it again. Is there any chance that you faced the same issue before, that your system only recognize your Card for couple of minutes at the beginning and not after that?
Also, I'm using Ubuntu 18.04 LTS. Please let me know if any other information could be useful.
I am using centos7 and Inatel PAC aria10gx.
I have expericed several times losting PAC in system, you can just input fpgainfo fme,and if there was no FPGA card information come out, you might know server lost fpga card for some reason.
one time was I downloaded a sof file through the signaltap, which is not supposed to do in that way, and I lost fpga and server restarted automatically,after system on,PAC card came back as normal as usual.
If you card can work for several minutes and lost recongnition, I think might caused by BMC ,which is used to monitor the current loading,temperature and some working parametters ,if some paratetters vaolated the limitation,BMS will shut the FPGA card from futher damaging.
Thanks, It could be case with BMC since it is a normal workstation and it may be over heated by the FPGA. The only doubt I have is that, if its because of BMC, why the problem could be fix temporary by shutting down the system and start immediately, but it could not be fixed by rebooting the system.
Is there any other way, I could check to make sure that the problem is from BMC ?
( also my board is the same PAC aria10gx)
Once server power on, the static region is loaded and PR region is empty I think.
Without loading gbs file into PR region, my server has no any issue no matter how long it runs.
Now I can successfully run many examples like I mentioned before without issue.
I can download different gbs files into PR region and perform tests.
If you are facing common BMC issue, you would see it through all other examples.
You can check if your version of FME and BMC is following the requirement in user guide, also you can check if your serer is validated or not based on user guide.
you might start to run nlb_mode_3 and dma_afu mentioned in the document Intel Acceleration Stack Quick Start.pdf.
different example draw different current loading, you might see different result.
I am still learning it, and still got so many other technical issues to solve, you can find my question in this field, hope my feedback can give you some idear.