Relatively beginner overclocker here. I have a 990fx Sabertooth and a Phenom II 1100t running at 3.84ghz cooled with a Xigmatek HDT-S1283 and Arctic Silver 5 (bought a little over a year ago) in a Rosewill Ranger case.
Attached is a picture of Prime95 running for an hour with CPU-Z and HWmonitor running. I maxed out at 60c at the cores and 64c on the main CPU temp. Are these temps okay? From what I've read, they seem a bit high. If so, any ideas why? The ambient temperature is about high 70s-low 80s (fahrenheit) today. Is my AS5 too old, or maybe I just did a bad job applying it?
Thanks. I think I've given the paste enough time to cure (a handful of P95s and a couple movies on handbrake), but I'll give it some more time. If nothing else, I'll buy some new thermal paste, probably MX-2, and try it again...I'm not totally confident I applied the AS right. HDTs are tricky to get right :-/
If worse comes to worse, I may just get another heatsink. Any suggestions on that front?
Well, with the 212 EVO suddenly pushed up to $45, I'm now recommending the $50 Scythe Mugen 3. I doubt a 212+ will do better than yours.
Do you think the Scythe would make a big enough difference to justify the $50 price tag?
/edit/ Also, I've heard good things about the Noctua NH-D14. Can you recommend that, or are there better ones in the $85 price range?
The Noctua's good, as are the Thermalright Silver Arrow and Archon and the Corsair H100 and this monstrous thing. I think the Mugen 3's the right choice, but otherwise the Archon would probably the best next step. What case do you have?
Rosewill Ranger, but I have an Antec 300 I can switch it out with...I'm on the fence about which case I like better, so I'm already considering it.
You are both forgetting about the faulty sensor in the x6's. The actual temperature is 10*C higher than read. So you are actually hitting 70*C. That could could be 5-8*C higher if you try testing with linX at a high memory usage (what I mentioned was the temperature difference I got when using 14GB of memory for the linX test)
About the stability, running prime95 will give you a stable system in 90% of use cases but if you really care about stability, you'll have to run prime95 for 24 hours, linX for say 8hour and then run the folding@ home project for 24 hours on all cores.
The tests: prime95 is good for basic testing of temperature and stability linX will push the CPU temperature higher than prime95 folding@home a4 will find instabilities where others won't.
I've had instances where my overclock has been 100% stable on prime95 but not on linX. I've also had stable on both but folding@home has still given me errors. Gaming you should be fine but if you are doing something calculation intensive (ie folding) you might want to do more stability testing.
ps. I've gotten the MX-2 on sale a lot and really like it.
It's very difficult to find the maximum heatsink heights for cases that don't have it on their manufacturers' websites. I wouldn't risk the Archon not fitting; honestly, the Mugen 3 will be fine.
enzo matrix said:
You are both forgetting about the faulty sensor in the x6's. The actual temperature is 10*C higher than read.
I haven't heard about that. Source?
Best answer selected by Inane_Asylum.
I know this answer is solved but I have the same Model of a cpu and what voltage are you running that at 1.45V should be the max. I have mine pretty low and have a Nice stable OC of 3.9Ghz.