Oh I know real temps are going to be far lower, but after reading up on everything I've been amiss on since Sandy first hit the shelves I'm being a bit more cautious than usual. And by bit, I mean an actual bit.
I'm just not comfortable with seeing the temperatures so close to the TMax. And yes P95 is boring to watch. Which is exactly why I always run it when I'm off to sleep.
lol, I'm actually doing 'real world tests' first and maybe I'll run it overnight.
I hit 104c degrees on a core during IBT - hows that for far too high?
HOAH! I hit 99C on my 2550K and shut 'er down mate. Of course you can put a pot on it for breakfast. BTW, I've has some setups pass IBT great and the BSOD in a few seconds or 7-8 minutes in P95.
You're a madman. And I'm going to be mad right alongside you, because you've inspired me to run the setup again. :)
That's pretty much why I don't put stock in IBT anymore; you can either pass IBT and fail P95, or pass P95 and fail IBT - I've never heard of someone passing both. And if they have... witchcraft. lol
I used to shut down 90c+ on SB. Luckily, ivy can handle a tad more. But at these temps, I'm going to be limiting my stress testing...
I have a 14c discrepancy between my coolest and hottest core, so I'll have to re-paste. 10c is one thing - above that and it's a botched application.
Real world, it tends to max out around the mid 60's. Still a good 10c hotter than real world 2500K temps. But since it's IVY with HT, it's not surprising. My i7 920 added about 10c by turning HT on.
It's gonna be interesting come summer time though
Just when I finally scratched the idea of watercooling because 'air was good enough, it's suddenly starting to look like a very good idea again
It does seem to idle cooler though which is one saving grace.
Muhaha - mad men unite!
My theory on IBT and Prime. If you set your Vcore by IBT you will fail Prime. If you watch the Vcore, When IBT starts it is higher vcore than when Prime starts. They both fluctuate some, but since IBT starts higher it can drop and still pass. Since Prime starts lower, when it drops you fail because it drops too low. You will get much higher temps with IBT, but if you set vcore with prime, you can pass both. I have never been able to pass Prime and fail IBT. It's just too easy to use IBT to set vcore, it finishes so quickly. But, I have passed IBT on max and failed prime. Those high vcores you guys are getting, you won't get in real world apps. At least I didn't with a 3770K, my real world apps were the same temps as my 2600K. I found the 2600K much harder to OC than the 2500K. It will do it, but you have to make it do it.
I'll find out tomorrow when my 3570K is delivered.
Interesting theory there, essenbe.
4.6 @ 1.365. Going on 7 hours strong, will stop the test at 8. And then it's for 4.7.
Not my current 3DM11 score by the way. It's just there for added effect. lol
I think it is just a quirk. My 2550K now has a 16C differential but three of the core are within 5C of each other. I took the time to lap both the CPU and cooler sink and I saw no improvement under full load. Idle is 3C-5C cooler.
As a note, when I removed the CPU cooler the paste was much in the center of the CPU and very thin on the edges. That means the CPU integrated heat sink is cup-shaped. I found this on two AMD chips as well.
I agree with your theory Steve, I have passed P95 and IBT at 4400 and 4600 but haven't tried it any higher. My vCore will go to 1.448 in IBT and about 1.400 in P95. Then after Prime runs the damn thing will BSOD sometimes....at the desktop.
Congrats
****
I just got my 3770K P95 12hr stable @ 4.7Ghz
Yes, those peak temps areThank flap they aren't 24/7 temps.
Cpu-Z
Settings that worked for me:
Now I'll see if I can lower the volts. It wasn't as much of as fight that I thought it would be to hit 4.7ghz. I'm quite pleased. This will do nicely for 24/7 settings.
