Page 1 of 1

New Multiforcer display: Feedback?

PostPosted: Sun Nov 21, 2010 7:33 pm
by Bitweasil
Playing with a new curses display module for the multiforcer.

Click the image to get rid of the scroll bars.

multiforcer-0_80_beta.png
multiforcer-0_80_beta.png (50.96 KiB) Viewed 4768 times


Yes, this does correctly imply that it will support multiGPU and CPUs ;-)

Feedback on the general layout? It will probably get slightly tweaked as I find out I need to add new things to display, but the general concept is pretty solid at this point.

Re: New Multiforcer display: Feedback?

PostPosted: Sun Nov 21, 2010 9:53 pm
by Bitweasil
Actual run data...

multiforcer-running.png
multiforcer-running.png (56.01 KiB) Viewed 4766 times


Now to get it faster. :D Fermi tuning here I come!

Re: New Multiforcer display: Feedback?

PostPosted: Sun Nov 21, 2010 11:09 pm
by blazer
i like how you can easily see the different stats. Will you add any sort of temperature sensing

Re: New Multiforcer display: Feedback?

PostPosted: Sun Nov 21, 2010 11:33 pm
by Bitweasil
blazer wrote:i like how you can easily see the different stats. Will you add any sort of temperature sensing


Looks like I might be able to wrap `nvidia-smi` to get GPU temps.

Would that be useful?

Re: New Multiforcer display: Feedback?

PostPosted: Mon Nov 22, 2010 6:55 pm
by Rolf
I like the new look.

Re: New Multiforcer display: Feedback?

PostPosted: Mon Nov 22, 2010 10:50 pm
by hackajar
Aside from debugging, is there a need for the WU count? Perhaps that space can be used for other statistics (Like Fan Speed, Core Temp, etc.)

Re: New Multiforcer display: Feedback?

PostPosted: Mon Nov 22, 2010 11:08 pm
by Bitweasil
hackajar wrote:Aside from debugging, is there a need for the WU count? Perhaps that space can be used for other statistics (Like Fan Speed, Core Temp, etc.)


The "WU Rate: " display?

That's really just a scrolling status display. Right now, all it's displaying during cracking is the rate calculated by the workunit module based on how long things have been running and how much work has been done. It's a bit below the displayed rate because the workunits are initialized before the data is copied to the GPU, so there's a few second delay. I could probably drop display of that data.

It sounds like "GPU core temp" is of interest to several people... I suppose I can wrap the nvidia-smi utility on Linux & get that data. Not sure how to do it on Windows, I'll get there when I get there. :)

Alternately, I could display the GPU temp by each GPU indicator.

Code: Select all
0: GPU (90C): 773.0M/s


Would something like that be useful?

Eventually, I'll have some overlay windows that can display useful data too. That was my intention for temperatures, but if they're of great interest, I can put them in the main display window.