[[SUPPORT] blakeblackshear - Frigate]

6 minutes ago, ich777 said:

Why not? I don't have that issue at all, or at least doesn't had it when I was running the TPU on my Server.

Would there be a command to lower the frequency that it runs at? I would like to see if i can fix this as sending it back is a pain and they are out of stock of these until the new year.

[[SUPPORT] blakeblackshear - Frigate]

On 8/9/2022 at 8:55 AM, ich777 said:

Please keep me update if it works after doing that and if the temperatures are stable.

OK after disabling the temp shutdown it still fails exactly when the txt files says -89.70C, so disabling the shutdown still has no effect on this. I have talked to the company that I bought it from and the suspect a faulty coral, but now I'm no so sure as @ssjucrono has the same fault as me? both cant have the same fault surly? 

[[SUPPORT] blakeblackshear - Frigate]

18 hours ago, ich777 said:

You can try to disable the shutdown from the module but be aware that you have to monitor the temperature on your own since even Google mentions in their documentation that the TPUs can be a fire hazard...

 

It is also possible that your temperature sensor on the TPU is defective.

 

To disable the shutdown execute the following line from an Unraid terminal:

sed -i "/shutdown_en0=/c\shutdown_en0=0" /boot/config/plugins/coral-driver/settings.cfg

and restart your server afterwards.

I have given this a go, I have also ordered the USB version and was wondering if I could use both at the same time?

[[SUPPORT] blakeblackshear - Frigate]

I have an issue with a coral TPU (Mini PCIe) that's on a Mini PCIe to PCI-E card plugged into a PCI Express slot. It will randomly shut down  and I need to restart the server to get it back, I used the script for temperature and it's not getting hot (I have put a heatsink and fan on the TPU). Here is the readings from the script :

 

2022-08-07 12:36:09 Coral Temp: 40.30C
2022-08-07 12:36:24 Coral Temp: 41.05C
2022-08-07 12:36:39 Coral Temp: 40.80C
2022-08-07 12:36:54 Coral Temp: 41.30C
2022-08-07 12:37:09 Coral Temp: 40.80C
2022-08-07 12:37:24 Coral Temp: 40.55C
2022-08-07 12:37:39 Coral Temp: 40.80C
2022-08-07 12:37:54 Coral Temp: 40.80C
2022-08-07 12:38:09 Coral Temp: 41.55C
2022-08-07 12:38:24 Coral Temp: 40.80C
2022-08-07 12:38:39 Coral Temp: 41.05C
2022-08-07 12:38:54 Coral Temp: 41.05C
2022-08-07 12:39:09 Coral Temp: 40.55C
2022-08-07 12:39:24 Coral Temp: 41.80C
2022-08-07 12:39:39 Coral Temp: 41.05C
2022-08-07 12:39:54 Coral Temp: 40.80C
2022-08-07 12:40:09 Coral Temp: -89.70C
2022-08-07 12:40:24 Coral Temp: -25.70C
2022-08-07 12:40:39 Coral Temp: 40.30C
2022-08-07 12:40:54 Coral Temp: -89.70C
2022-08-07 12:41:09 Coral Temp: -89.70C
2022-08-07 12:41:24 Coral Temp: -89.70C
2022-08-07 12:41:39 Coral Temp: -89.70C
2022-08-07 12:41:54 Coral Temp: -89.70C
2022-08-07 12:42:09 Coral Temp: -89.70C
2022-08-07 12:42:24 Coral Temp: -89.70C
2022-08-07 12:42:39 Coral Temp: -89.70C
2022-08-07 12:42:54 Coral Temp: -89.70C
2022-08-07 12:43:09 Coral Temp: -89.70C
2022-08-07 12:43:24 Coral Temp: -89.70C
2022-08-07 12:43:39 Coral Temp: -89.70C
2022-08-07 12:43:54 Coral Temp: -89.70C
2022-08-07 12:44:09 Coral Temp: -89.70C
2022-08-07 12:44:24 Coral Temp: -89.70C
2022-08-07 12:44:39 Coral Temp: -89.70C
2022-08-07 12:44:54 Coral Temp: -89.70C
2022-08-07 12:45:09 Coral Temp: -89.70C
2022-08-07 12:45:24 Coral Temp: -89.70C
2022-08-07 12:45:39 Coral Temp: -89.70C
2022-08-07 12:45:54 Coral Temp: -89.70C
2022-08-07 12:46:09 Coral Temp: -89.70C

 

I'm assuming the reason for the negative temperature is because of the shutdown, I cannot see any fault in the logs other than :

 

Aug  7 12:49:40 Tower kernel: apex 0000:05:00.0: Apex performance not throttled due to temperature
Aug  7 12:49:45 Tower kernel: apex 0000:05:00.0: Apex performance not throttled due to temperature
Aug  7 12:49:50 Tower kernel: apex 0000:05:00.0: Apex performance not throttled due to temperature
Aug  7 12:49:55 Tower kernel: apex 0000:05:00.0: Apex performance not throttled due to temperature
Aug  7 12:50:00 Tower kernel: apex 0000:05:00.0: Apex performance not throttled due to temperature

 

Over and over until the server is reset. Any idea what's happening?

 

Thanks.