measuring device temperature

Hi,

Does anyone know how to measure the device temperature? Thanks.

well speedfan does a good job at monitoring your CPU temps but for GPU temps i would use EVGA precision as a nice monitoring tool but that works just for the graphics cards

If you are talking about checking the temps of the GPU then several pieces of software will do that without issue. Speedfan, EVGA precision tuner, realtemp and several others.

EVGA precison tunes is nice because it allows you to control the fan speed of the GPU.

http://downloads.guru3d.com/EVGA-Precision-1.8.1_d2403.html

Oh you beat me SLAY

well you did give a little better explanation on what the softwares do lol

Of course, those are all windows applications… on linux you can use nvidia-smi or nvidia-settings to see the current temperature.

Mainly I just poke it.

(nvidia-smi is probably the right answer, but there’s no good equivalent for Windows that gives you robust logging support)

I tried that last week on a running cluster node I was troubleshooting. The results weren’t pretty. Blood was shed…

where I can download it from?

Download which one? Speed fan, EVGA precision etc. or nvidia smi?

nvidia-smi, i.e. the linux based software.

I believe it’s bundled with their drivers (someone correct me if I’m wrong) so:

http://www.nvidia.com/Download/index.aspx?lang=en-us

Yes it is bundled with the Nvidia linux drivers

nvidia-smi is bundled with any of the recent drivers with cuda support, so you should just be able to do this:

avid@cuda:~$ nvidia-smi  -lsa

==============NVSMI LOG==============

Timestamp			: Wed Nov  4 22:54:17 2009

GPU 0:

	Product Name		: GeForce GTX 275

	PCI ID			: 5e610de

	Temperature		: 44 C

GPU 1:

	Product Name		: GeForce GTX 275

	PCI ID			: 5e610de

	Temperature		: 45 C

cheers for the clarification :thumbup:

Any ideas for troubleshooting of the following?

==============NVSMI LOG==============

Timestamp : Thu Nov 5 12:56:06 2009

GPU 0:
Product Name : GeForce GTX 280
PCI ID : 5e110de
Failed to read GPU temperature!
Temperature : 0 C

Weird. Try updating to the latest driver version, maybe?

we updated to cuda 2.3 - it didn’t work.

i’m having the same problem even with 2.3. Could it be that I’m using the workstation remotely through ssh?

GPU 0:
Product Name : Tesla C1060
PCI ID : 5e710de
Failed to read GPU temperature!
Temperature : 0 C
GPU 1:
Product Name : Quadro NVS 295
PCI ID : 6fd10de
Failed to read GPU temperature!
Temperature : 0 C

I think 190.18 was broken for standalone cards but newer drivers work.