PCIe Fatal Error

Hi, I have a Z820 workstation with 2 nvidia cards (Quadro 6000 and Tesla C2075) and I have a common problem:

The monitor turns off and soon returned saying “Blast graph stopped working and recovered” …

setting “Maximum Performance” in graphics settings and OK … and a few days ago appeared black screen saying:

"928-PCIe Fatal Error.
6 Slot PCIe error detected
Completion timeout
"

QUADRO card:
-is in slot 5 (PCI v2.0)
-Speed ​​is 5 GT / s

Testla card
-is in slot 6 (PCI v3.0)
-Speed ​​is 2.5 GT / s

thanks.

Your post is a bit vague, what O/S? What programs are you running on the GPU when the errors occur? Do you have the latest drivers installed? etc. Post more details of a possible root cause.

In addition to ensuring that the latest NVIDIA drivers are installed, I would suggest the following:

(1) Make sure the latest system BIOS is installed
(2) Check that both GPUs are properly seated and secured in their slots
(3) Check that all GPU power connectors are fully inserted

Is the wattage of the power supply in this machine sufficient to drive two GPUs that together could draw up to 419 watts (according to NVIDIA specifications available online the Quadro6000 is a 204W part, and the Tesla C2075 is a 215W part) ?

As a separate note, we will be posting a notice on these forum in a week that warns users about Windows Server 2008 installation with a fibre channel HBA in Slot 5. The R900 system may produce a PCIe Fatal Error B0 D1 F0, STOP Error 124 when installing Windows Server 2008 if a fibre channel HBA is in slot 5. This is not related to your error, however, just want to give you upfront notice in case you run into this. The fix for this error will be a hotfix from Microsoft.