Quick Questions about my CPU Temps / GPU dummy plug

[rjbelans]

I'm new to F@H (3 days now). First of all, this is a great idea, using as much computing power as possible that is out there which would othwise go unused.

As I do not want to get into SMP right now, I am currently running 4 CPU clients and 2 GPU clients (will try for 4 GPU soon). On my Core i5, the core usage is 100% on all cores, and the current Core Temps for all 4 are hovering between 64C and 67C.

My questions:

1. Am I right to believe hat this is fine to run 24/7 or should I reduce the usage a bit to reduce temps?

2. At what temps should I start reducing the usage? (Assuming that the above temps are ok)



Sorry, no (Deleted: scripts or) system info posting right now. Maybe soon. Couldn't figure out how to make strikout font, so made it red.

[John Naylor]

1) those should be fine 24/7... many people run at hotter temps than that (myself included) with no adverse side effects.

2) I get the jitters around 75C, but some chips can take more... I certainly wouldn't recommend maintaining more than 80 however (again, some do, but its probably not a bad cutoff point).

[rjbelans]

Thank you.

I wasn't running around trying to find an ice pack to put on it just yet since I had read somewhere about cores being able to handle spikes of around 100C. I just didn't want to get into trouble with the 24/7 running.

I now feel more comfortable with trying to get more GPU clients running, once I make a couple of the adapters (dongle, dummy plug) to fool Win 7 into thinking all of the cores have a monitor attached.

Time to work on getting a script or two for my sig now.

[weedacres]

What kind of temps are your current gpu's running?
Adding 2 more gpu's will generate a lot more heat in your case. Good air flow through your case is the key to keeping temps under control.

[bruce]

1) those should be fine 24/7... many people run at hotter temps than that (myself included) with no adverse side effects.

2) I get the jitters around 75C, but some chips can take more... I certainly wouldn't recommend maintaining more than 80 however (again, some do, but its probably not a bad cutoff point).

3) No matter what cutoff point you select, GPUs are expected to run at higher temperatures than CPUs.

[rjbelans]

What kind of temps are your current gpu's running?
Adding 2 more gpu's will generate a lot more heat in your case. Good air flow through your case is the key to keeping temps under control.

The GPU Temps are (were) about 50C. I say were because one seemed to go on the fritz. Monitor started flickering, so pulled one card and all was good on both ports, then tried the other one alone and one port was good, but the other showed while as yellow right from POST and BIOS screens. Weird, but at least I've had it a very short time, it's going back to newegg for a new one. (Side note: They actually sent me a UPS label to cover the cost of the RMA shipping. That's a new one for me.)

Only running one GPU client right now. Is the only way for me to run two by having a monitor or "dummy plug" insterted into the second port? When both cards were in and running in SLI, I could run two GPU clients with only one monitor plugged in. (I am running Win7) When I do get my second card back in, will I need three dummy plugs or maybe just one?

[weedacres]

Is the only way for me to run two by having a monitor or "dummy plug" insterted into the second port? When both cards were in and running in SLI, I could run two GPU clients with only one monitor plugged in. (I am running Win7) When I do get my second card back in, will I need three dummy plugs or maybe just one?

When I was running Win 7 (have now upgraded to XP ), I plugged 1 monitor into each card, not each port. I was not using SLI so can't speak to that. In my case I use a KVM switch which does the same thing as the dummy plug. And of course extended the desktop to each card.

[bruce]

It's possible to run more than one GPU client with only one GPU active (whether or not a second GPU is installed) but they'll compete with each other with the resources for that GPU. FAH won't use SLI; a second device needs to be recognized by the OS. If you have an OS which requires a dummy plug or monitor to be connected to each GPU so that you can extend the desktop to the second GPU and so it stays enabled, you'll need to do that. (If your OS doesn't care, then neither does FAH.)