[QUOTE=James Heinrich;253701]How difficult would it be to put in some code (at least for the Windows version) to detect keyboard/mouse activity? What about running processes?
Dreaming for a second, what I'd like to see (configurable in mfkatc.ini):[list][*]if keyboard/mouse active within last <x> seconds, throttle mfaktc to <y>% (0-100)[*]if <program> active, throttle mfaktc to <y>% (0-100)[/list]This would let me run mfaktc a lot more, since I just don't run it most of the time (my system is nigh unusable when running), and I often forget to start it when I leave the computer for an extended amount of time. If it could just run in the background slowly (or paused, if I'm running something specific), and then kick into high gear when I'm gone, that would be fantastic. :smile:[/QUOTE]

Did you try to force SievePrimes in mfaktc.ini to the lowest/highest possible value? It should keep your GPU idle for more time while waiting for the completion of each thread.

Luigi

[QUOTE=ET_;253711]Did you try to force SievePrimes in mfaktc.ini to the lowest/highest possible value? It should keep your GPU idle for more time while waiting for the completion of each thread.[/QUOTE]If I force it to 5000 (lowest allowed value, I believe), it's very very slightly better than at 100000, but not really. It seems a fast(ish) CPU and and old slow(ish) GPU are a bad combination :no:

Here's where reduced threads come in handy.
Set them to 32, instead of 256, and desktop should be more responsive.

[QUOTE=Karl M Johnson;253806]Here's where reduced threads come in handy.
Set them to 32, instead of 256, and desktop should be more responsive.[/QUOTE]Forgive my ignorance, but I can't find that setting in [i]mfaktc.ini[/i]?

James, the number of threads (per block) is a compiletime option. But you can try to lower NumStreams in mfaktc.ini.

Oliver

[QUOTE=TheJudger;253828]James, the number of threads (per block) is a compiletime option. But you can try to lower NumStreams in mfaktc.ini.[/QUOTE]Lowering to 1 did improve things considerably. But it also cut my rate from 38M/s to 18M/s. Going to 2 streams is still noticeably better than 3, and only cut my rate from 38M/s to 37M/s, so I can live with that.

I'd still like to see a [i]PauseWhileRunning[/i] option though. :smile:

Could you make your system call it as your screen saver?

Chris K

[QUOTE=chris2be8;253840]Could you make your system call it as your screen saver?[/QUOTE]
Haven't tried [URL="http://www.softpedia.com/get/Desktop-Enhancements/Screensavers/Screen-Launcher.shtml"]this[/URL] software but it seems to offer a (freeware) solution (on Windows). I'd be interested if it works or if you find another solution.

A single instance of mfaktc uses about 53% of the GPU (and a full core, ie 25%), and I can run two instances, to increase the overall M/sec, pushing the GPU to 100%, but reducing the M/sec for each instance.

If the second instance of mfaktc is started afterwards, how does it know how much GPU resources to request? Does it equally balance the resources used by each program? Or even between instances of CudaLucas and mfaktc for that matter?

Thanks.

To fully use mfaktc here, I need to launch 3 instances of it.
Short answer - CPU bottleneck.

[QUOTE=Karl M Johnson;254085]To fully use mfaktc here, I need to launch 3 instances of it.
Short answer - CPU bottleneck.[/QUOTE]
Have you tested 4 instances for total throughput? 3 might saturate the GPU only with a very low sieveprimes bound which would be more optimal when raised. Maybe try 4 mfaktc instances with something else like llr sharing one of the cores.