I'm having trouble building code (both example/lab from Intel and personal) while targeting for Stratix 10 FPGAs on the DevCloud. I'm issuing a .sh file that looks like the following (based on a oneAPI workshop lab):
source /opt/intel/oneapi/setvars.sh > /dev/null 2>&1
rm -Rf bin
cp -pR bin-original bin
dpcpp -fintelfpga ./source/fpga_mst.cpp -Xshardware -Xsboard=intel_s10sx_pac:pac_s10 -Xsprofile -o ./bin/fpga_mst.fpga
And shortly after issuing it (~20 seconds) I am getting this error in return:
Can't opendir /glob/development-tools/versions/oneapi/2022.1.2/oneapi/intelfpgadpcpp/2022.1.0/board/de10_agilex: Permission denied at /glob/development-tools/versions/oneapi/2022.1.2/oneapi/compiler/2022.0.2/linux/lib/oclfpga/share/lib/perl/acl/Common.pm line 717.
dpcpp: error: fpga compiler command failed with exit code 13 (use -v to see invocation)
Removing the compilation flag "-Xsboard=intel_s10sx_pac:pac_s10" allows the synthesis to proceed, however, targeting the default FPGA of Arria 10. I think the issue is caused by the tools checking the board directory for a valid S10 configuration, but finding the agileX board instead, which I do not have permission for, and failing.
Any insight on how to solve this error would be much appreciated, as the issue is also experienced by my peers trying to build for Stratix 10 boards as well.
Thank you for posting in Intel community forum and hope all is well.
My guess would be the same as you mention, would you be able to let us know in which nodes within the Devcloud you are building the code on for us to check further?
Hope to hear from you soon.
Thank you for your response, it appears that the issue has been resolved over night and I am now able to build for the Stratix 10 FPGAs.
I've encountered a new problem, however, as it appears that I cannot run any programs on the Stratix 10 devices. I'll attach a text file displaying the various error messages and outputs from the aocl commands attempted. Any insight that you can provide on them would be much appreciated!
Noted on the updates, it seems that there might be some platform instability which has ben resolved.
We have notice the same problem suspecting hardware connection on our side too in the mention node too and have escalated this to the relevant team.
Will get back to you once we get an updates on this.
Thank you for the patients.
Updates on the 'Error enumerating AFCs: not found', it seems to be an hardware anomaly, and has been attended and verified fixed from our end.
Please do try from your end and let us know if issues is persisting.
Thank you for the understanding.
Great! Good to know that it is working now and apologies for the inconvenient cause, with no further clarification on this thread, it will be transitioned to community support for further help on doubts in this thread, where we will no longer monitor this thread.
Thank you for the questions and as always pleasure having you here.