Temp question/Should I OC?

Hey guys,

Just put my new rig together, but unfortunately, it was last minute and couldn't afford water cooling. Instead I've got a Mugen 2 cooler on my CPU and my idle temps are ~34C (the cores are each ~25C if that makes a difference). My CPU is the AMD 1090T BE. Stock speed is 3.2GHz.

My question is: with my current fan/heatsink and knowing what my idle temps are, do you think I could get away with an OC? I'd like to get it between 3.6-4GHz. I want to do it more for the experience than anything.

I know I'll be asked for load temps which is why I'm going to run Orthos over night (Blend/Priority - 6).

Anyone have experience with the cpu/heatsink combo?

Most people with Phenom II CPUs are comfortable with load temps in the mid 50 C range. You should be able to do some OC'ing and limit your frequency and voltages based on load temps.

See Dolk's Phenom II OC'ing thread for more detail.

http://www.overclockers.com/forums/showthread.php?t=596023

Thanks I'll check it out. I rand OCCT for two hours and didn't even hit 40C

OC'd to 3.4GHz and ran OCCT with high priority for another 2 hours. Still didn't hit 40C. Gonna go up to 3.6GHz and see if I'm still stable.

Decided to try Intelburn with my 3.4GHz OC and my temps are running at around 48C. I know I shouldn't run over 50C so I'm not sure if I can do 3.6GHz.

I think most cpu temps can go up to 100c before it shuts down

I'm on one a h50. Corsair and load twmps I get is max 80c lol overclocked at 4.0

Rotary said:

I think most cpu temps can go up to 100c before it shuts down

Click to expand...

He has an AMD man....

So it can't get that hot? Intel runs hotter I guess

Well, Im not sure where its thermal cutoff is, but I know stablity suffers greatly around 55-60C or more so there isnt a point to go over that.

Bob, gonna beat you to the punch here:

I think (via bob) that the amd temp sensors are just placed differently. They probably max about the same at the cores, but amd just read em differently than intel. Intel looks for the hottest spot, amd puts em where there's room.

@shade - don't bother with a 2 hr test at each oc, 10 mins linx (for me at least...I'm on an intel though, there is creep on amds) should be enough for a quick stabililty test to get you where you're going faster. once you reach you goal or where it's not linx stable, back it off some, then go for your 8-24hr prime run.

m0r7if3r said:

Bob, gonna beat you to the punch here:

I think (via bob) that the amd temp sensors are just placed differently. They probably max about the same at the cores, but amd just read em differently than intel. Intel looks for the hottest spot, amd puts em where there's room..

Click to expand...

Whoa really? Never heard that one... and its been that way for a couple generations I would imagine b/c since the S754 days, they say 50-55C is fine but more than that can hurt stability...

That's my understanding. I know that Intel scatters temp sensors all over the place, and then in post-production tests them all and picks one per core that reads the hottest. AMD meanwhile only has one temp sensor for the whole die. I think it's probably because AMD simply cannot "waste" the space on un-used sensors, they're in the business of packing as much power into a tiny die as possible. Intel on the other hand is more focused on packing the most power into a given thermal profile. (It's much like ATI vs Nvidia. ATI uses small highly efficient dies, nvidia uses much larger but less efficient. Both strategies work.)

55*c is generally where stability goes south.
I think thermal spec is 62*c, and total shutdown is 72*c.

Aim for 55*c max, anything that temp or under is fine.
Be sure to test temps for at least ten minutes, no less! Really half an hour is better.
I'd recommend OCCT or Prime95 rather then IBT as IBT gives the cpu short breaks between tests and temps drop quite a bit. Not especially good for testing max temps.

Bobnova said:

I'd recommend OCCT or Prime95 rather then IBT as IBT gives the cpu short breaks between tests and temps drop quite a bit. Not especially good for testing max temps.

Click to expand...

I've found, despite the breaks, that linpack variants are actually able to get my cpu temp higher than anything else (and I've tried em all)...I suppose the huge load is so much more intense than anything else that the breaks don't matter, can't really say why. That said, I use linx for quick and dirty (because of aforementioned heat generation capability) and prime for long and slow (linx tends to have issues manifest quickly, so 20-30 mins on linx and you'll draw anything that it'll throw)...just my method, there are MANY others that are completely valid.

Sounds good to me. I haven't done a lot of testing of different methods on AMD.

Bobnova said:

Sounds good to me. I haven't done a lot of testing of different methods on AMD.

Click to expand...

I've never oc'd an amd (got one that i'm gonna do soon tho)...this method is based off my i7

OP-

OCCT and Intelburn Maximum, will give you a quick stability/temp check. Even with the Mugen 2 I would expect your OC'd CPU temps to get into the 50's if your idle temp is ~34 C. AMD has a Max operating temp of 62 C listed for the 1090T so full load Intelburn Maximum test temps in the 50 C range are no problem at all.

http://products.amd.com/en-us/Deskt...2=&f3=&f4=&f5=&f6=&f7=&f8=&f9=&f10=&f11=&f12=