10-04-2020 02:08 PM
I've stumbled upon a critical problem with my two Alveo U280 cards.
My server machine is unable to detect my two Alveo U280 cards since I programmed them with my custom RTL example, which is just a QDMA IP with loopback (this is not the first time I've programmed the Alveo U280 via custom RTL flow, it has worked fine before).
I programmed one card with the generated .bit file from Vivado, and programmed the other with a .mcs file I created from the .bit file.
After doing a cold reboot, neither card showed up under lspci or the Vivado Hardware Manager. However, you can see in the image below that the Vivado Hardware Manager detects both targets, but no devices.
I now have no way of reverting the card to factory settings via xbutil or the Vivado Hardware Manger because neither device is detected. The QDMA IP interfaces with PCIe, and therefore, may have tampered with the board's PCIe interface, however, I'm not sure why the board target isn't detected in the Hardware Manger.
Below are the steps I've taken to resolve the issue (none have been successful):
1. Reinstalling cable drivers
2. Multiple cold reboots
3. Made sure CAT TRIP pin is driven low in my constraints file.
4. Tried different JTAG frequencies to see if the device shows up, but no luck.
I've invested a lot of time in these using and developing on these Alveo cards, so it's imperative that I get them back up and running ASAP.
I would truly appreciate any help I can get! I'm fresh out of ideas
Thank you, in advance!
10-05-2020 05:31 PM
A quick update, that one of my cards came back with a blue LED after yet another cold reboot. However, my second card (shown in the image below on the right) is still displaying a red LED.
I've tried the fixes that were listed in the "Known Issues" section of the Alveo getting started guide, but the issue still persists with my second Alveo card.
I would still very much appreciate you help.
Thank you, in advance!
10-11-2020 10:58 AM
Hi @hseyedro3 :
Try swapping the cards and verify if the second card (bad one) RED LED is still on.
Try removing the good card and plug the bad card in the server alone and verify.
If RED LED is not flickering and constantly ON, shut down the whole system, remove power cables, reinstall the card, and verify again.
If the issue "persists", you can request for RMA provided warranty seal is not broken. Note that you need to provide all relevant information acknowledging the issue if RMA request is filed. Refer https://www.xilinx.com/support/answers/72533.html for more information.
11-04-2020 11:10 PM - edited 11-04-2020 11:10 PM