When I get some free time I'm going to write a GPU temperature report plugin.
I haven't decided whether it's going to be an actual plugin or something I'm going to integrate into Core Temp (but will come as a separate DLL file).
Here's the preliminary test:
This only works on nVidia GPUs so far. I use nVidia's NVAPI library they provide on their developers' page.
I'll have to dig around and see if ATi offers something similar. As far as I know they don't, so if anybody has any information to share about it, please do.
The screenshot in the first post was just a proof of concept app to see if this was actually working.
I still haven't really thought about how it's going to be implemented, at all.
So I'll think about it in due time. I may do it the way you suggest, but the problem is that there may be up to 3 sensors per GPU, and there may be upto 8 or more GPUs in the system. So if/when I add this to shared memory, I'll have to think of something on how to allow access to this information to other programs.
I've uploaded this program, and am asking for some beta testers.
I'm pretty sure that the code works but I still want to make sure everything's correct.
Could you please run it and post a screenshot with the results?
This currently works on nVidia cards only.
Been a bit busy lately.
Not much progress on this front yet.
I need to figure out where to put GPU temp readings in the GUI, I may need to create another window to display the temps
If you need ATI drivers so you can see what makes them tick and how it displays the GPU Speed/GPU temps in ATI CCC (Catalyst Control Center), I would be more then happy to upload them for you so you can get the ATI aspect of it running.
I am just curious as to what language you use for programming your applications, and if you could give me a lead of direction in reading hardware temperatures etc. I have messed around allot with the windows WMI class and the temperature reading in it seem to fail badly.
Core Temp is written in C++. To gain direct access to hardware I'm using a kernel mode driver (ring0), since reading the MSRs or accessing the PCI bus requires access to ring0.
I've not messed with WMI as I had no need for it, so I can't really suggest anything to you in that department.
Would you have any knowledge or idea of reading ATI Graphics Card temperatures, Maybe there is a class file which can be used which is part of the catolist control center as it reads gpu temperature.
ATi has an API to read the temps but it's buggy and unreliable. The only reliable way of reading the temps would be doing it through the I2C bus, and that's a pain. It requires a lot of work, needs driver support (driver will have to be updated).
I don't yet know how to work with the I2C bus directly, so until I figure it out, this will be limited to nVidia only, unfortunately.
I'd be very happy if ATi implemented a similar API to nVidia's in this regard, nVidia makes it very simply and straight forward.
Yea true, i was using an Nvida card but upgraded to a ATI recently, ive still got some Nvidia software installed which won't allow me to uninstall without putting my old gfx in so dud.
What do you guys think?
The plug-in docks itself below Core Temp and it changes position together with the Core Temp window itself, so whether you move it, minimize it, go to mini-mode, etc, the Plugin window will be "dragged" together with the main window.
CPUID, APICID and Revision can all fit on a sigle line.
Then I would add a single GPU window, Vendor, Model and Memory on a single line. Every next lines show temp just like for CPU with Low, High. Don't know if you can read load with you plugin.
Not sure number of sensors is needed since you will display temp for every sensors so if 2 temps, 2 sensors.
