Case #: 00288368 - Unable to configure Arria 10 GX FPGA Development Kit for OpenCL

[aejjeh]

I am trying to configure our Intel Arria 10 GX FPGA Development Kit for use with the OpenCL SDK by following thefollowing guide. I am able to reach "Installing the OpenCL Runtime Driver - step 3". However, when I get to the point where I try to program the flash memory "Programming the Flash memory on the Intel Arria 10 GX FPGA Development Kit - step 1,b", aocl flash is unable to detect the device. This happened after I installed the Intel SDK for OpenCL Applications version 7.0.0  (note that we need the SDK because we also use GPUs in our development environment).

Note that even before I install the SDK for OpenCL Applications, I was having an issue configuring the Arria 10 Development Kit where when I get to the last step in the guide where it asks for a hard reboot of the machine, the board is not detected anymore by the OpenCL runtime after I perform the reboot. However this is a different issue which I think might not be relevent to the first part above. I do, however, feel that there might be a hardware issue with the board not being able to boot from the on-board flash. 

Please we need to be able to resolve the first issue ASAP since our research depends on it. The second issue can be tackled later.

[HRZ]

Well, your board might really be broken, after all. Maybe the PCI-E connection is not working correctly or something. Have you tried the board on another machine? In fact, can you post the details of the machine you are using with respect to the motherboard model and what other PCI-E devices you have on the motherboard?

One other thing I can think of is that if the A10 ref board has an on-board flash (it should but I have never used that board) that can be used to automatically program the FPGA after a cold boot, you can program that flash using the OpenCL image and then do a complete shutdown and start. Doing so can help make sure the motherboard is not cutting the PCI-E power and wiping the FPGA during soft reboot.

[SSFyTMT]

Hi HRZ - programming the on-board flash with an openCL image was the point of AN807 to configure the A10 GX board. But to do that it says it has to use the "aocl flash" utility - which means the board has to at least be recognized.

I was able to resolve the issue. The motherboard I was using had two PCIe edge cards in use - a 10GigE NIC and the accelerator board. The solution was to put the A10 GX board in the slot that was being used by the 10GigE NIC and vise versa. Again, a bittware accelerator board using an Arria 10 works fine in the original slot. I suspect this is some bug/incompatibility in the motherboard PCIe chipset and the .aocx image provided in the bringup folder of the a10_ref BSP. I found one other thread buried on stack overflow having a similar issue - if I can find it again I'll link it here.

While I can kick myself for not trying this sooner, this whole endeavor was an exercise in "just try stuff until it works" because there were limited handles to debug the issue. An Intel engineers ended up telling me no one had ever followed AN807 on Windows (even though Windows is supported) - which was likely a red herring. Then I was told it would likely on work on Redhad/CentOS and not Ubuntu - another red herring. I had a another Intel FPGA Accelerator board from bittware that is clearly based on the a10_ref BSP working in that slot - which pointed away from HW or chipset issues.

[MuhammadAr_U_Intel]

Hi @SSFyTMT​ 

If I understand correctly you manage to proceed with the Arria10 Devkit setup changing the PCIe slot on your motherboard. Was there any difference on both the PCIe slots ?

Are both of them PCIe Gen3 x 8 ?

--

Arslan

[MuhammadAr_U_Intel · Accepted solution]

Copying here a known issue with Ubuntu 16.04 (kernel 4.14 or newer), just in case anyone land here searching on this.

https://www.intel.com/content/altera-www/global/en_us/index/support/support-resources/knowledge-base/embedded/2018/why-does-aocl-diagnose-fail-to-find-device-on-ubuntu-16-04-5-lts.html

[HRZ]

Finally got access to my account again after over a week of being locked out...

Anyway, @SSFyTMT​ I am glad to hear that you finally managed to find the root cause of the problem.