I recently got the Tesla K80 and I installed it dual with GTX 1080. My motherboard (MSI Z-370 PRO) had old BIOS (2017). Installation went fine as I initially removed GTX 1080, booted up and installed Tesla drivers, then shutdown and installed GTX1080 back. Everything was working without any issues.
Then I updated my MB’s BIOS to the latest version and did the installation same as described above (removing GTX …etc.) and now I am getting this in nvidia-smi:
Wed Dec 02 19:21:07 2020
±----------------------------------------------------------------------------+
| NVIDIA-SMI 441.12 Driver Version: 441.12 CUDA Version: 10.2 |
|-------------------------------±---------------------±---------------------+
| GPU Name TCC/WDDM | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 1080 WDDM | 00000000:01:00.0 On | N/A |
| 20% 40C P0 34W / 215W | 630MiB / 8192MiB | 0% Default |
±------------------------------±---------------------±---------------------+
| 1 Tesla K80 TCC | 00000000:05:00.0 Off | N/A |
| N/A 34C P8 27W / 149W | 9MiB / 11448MiB | 0% Default |
±------------------------------±---------------------±---------------------+
| 2 Tesla K80 TCC | 00000000:06:00.0 Off | 0 |
| N/A 30C P8 30W / 149W | 9MiB / 11448MiB | 0% Default |
±------------------------------±---------------------±---------------------+±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1060 C+G …4.23.0_x64__8wekyb3d8bbwe\YourPhone.exe N/A |
| 0 1596 C+G Insufficient Permissions N/A |
| 0 2140 C+G …6)\Google\Chrome\Application\chrome.exe N/A |
| 0 2808 C+G …1.0_x64__8wekyb3d8bbwe\WinStore.App.exe N/A |
| 0 4716 C+G C:\Windows\explorer.exe N/A |
| 0 5208 C+G Insufficient Permissions N/A |
| 0 6900 C+G …11102.0_x64__8wekyb3d8bbwe\GameBar.exe N/A |
| 0 8008 C+G …w5n1h2txyewy\InputApp\TextInputHost.exe N/A |
| 0 9112 C+G …a\Local\Discord\app-0.0.308\Discord.exe N/A |
| 0 10264 C+G …5n1h2txyewy\StartMenuExperienceHost.exe N/A |
| 0 10860 C+G …dows.Search_cw5n1h2txyewy\SearchApp.exe N/A |
| 0 11504 C+G …osoft.LockApp_cw5n1h2txyewy\LockApp.exe N/A |
| 0 14008 C+G …0.92.0_x64__kzh8wxbdkxb8p\DCv2\DCv2.exe N/A |
| 0 14116 C+G …9.4.0_x64__8wekyb3d8bbwe\Calculator.exe N/A |
| 0 15340 C+G …10.2009.5.0_x64__8wekyb3d8bbwe\Time.exe N/A |
| 0 16428 C+G …1011.0_x64__8wekyb3d8bbwe\commsapps.exe N/A |
| 0 17232 C+G …mmersiveControlPanel\SystemSettings.exe N/A |
±----------------------------------------------------------------------------+
WARNING: infoROM is corrupted at gpu 0000:05:00.0
WARNING: infoROM is corrupted at gpu 0000:06:00.0
The only thing that changed was the BIOS update.
Also, when I check the NVIDIA Control Panel, one of the K80 GPUs (it displays two of them as usual) shows it does not support ECC. Before BIOS update both of them showed ECC support.
It is worth noting the old BIOS software did not recognize the K80 at all. It used to display “Unknown PCIe device”, but the updated BIOS shows “NVIDIA 3D Renderer”.
Testing on live applications showed no difference at all.
Should I be worried? Do you recommend I reflash the BIOS with the old firmware of 2017?