Did I just get lucky?

Toggle sidebar Toggle sidebar
G

Guest

Guest
I recently purchased a megalhelms cooler and a new antec 300 case to use with my i7 920. It's a c0 part which I have read are significantly worse overclockers than a d0 one. Now I can easilly hit 3.8ghz at stock voltage using either a 19x200 or a 20x190 setting. At 20x190 I am seeing idle temps of around 40c going up to 65c at full load under prime95. Thing is if I change to 19x200 idle temps go down to 28c (!) and full load is at 38-38c. Was I just lucky enough to get a particularly good CPU or is there some underlying reason why equivilant speeds on the same CPU would produce such different heat characteristics?

Mb is a gigabyte udr3.
HT is disabled.
 
Sort by date Sort by votes
Yea I realize that but is there any explination for the observed behaviour of the same speeds at different multiplier x bclk settings resulting in different levels of dissapated power?
 
Upvote 0 Downvote
Ok, seems its my bad. I was using a program called fan speed which apparantly has a known issue of reporting temps 10-15 C lower than they actually are. Using realtemp now which is reporting at 56C under full load. Guess thats still pretty nice for 1.2Ghz of free speed.
 
Upvote 0 Downvote
I'm still curious however about the observed temperature difference which is still present, albeit 10C higher, between the 19x200 and 20x190 settings.
 
Upvote 0 Downvote
i had to push the voltage up slightly to 1.25V after a fault running a prime95 workunit. It was one fault in about 7 hours run time but technically not stable in the end at stock. I am 2 hours into a repeat test now and everythings fine so far, temps are sitting at 57C on the highest core, lowest is at 55. I'm hoping the extra .05V will take care of the minor instability.
 
Upvote 0 Downvote
I'm still curious however about the observed temperature difference which is still present, albeit 10C higher, between the 19x200 and 20x190 settings.
Click to expand...

I was also curious about this, so I changed mine from 20x190 to 19x200, and I'm also seeing lower temps. Not a massive drop. Hottest core is 3 degrees cooler than before after 10 minutes of testing. I might keep it like this, or try to push it faster.

Sounds like you really did get lucky, though.
 
Upvote 0 Downvote
The reason I am curious about it is that from my understanding the skin effect should increase resistance and hence dissapated power as a function of frequency. It seems counter intuitive that an increase in bclk frequency should reduce heat given this knowledge.
 
Upvote 0 Downvote
i had to push the voltage up slightly to 1.25V after a fault running a prime95 workunit. It was one fault in about 7 hours run time but technically not stable in the end at stock. I am 2 hours into a repeat test now and everythings fine so far, temps are sitting at 57C on the highest core, lowest is at 55. I'm hoping the extra .05V will take care of the minor instability.
Click to expand...

Try intel burn test which is better than prime95. you think your system is stable until i test it with intel burn test. also, i take HT over extra mghz.
 
Upvote 0 Downvote
Yeah LINPACK is absurdly stressing, but Prime95.... OK
Still, I don't think at any given moment you will stress all your cores 100% for longer than say 15 minutes...
 
Upvote 0 Downvote


How about encoding videos with an app that supports multithreading properly? Linpack does give you the worst case scenario in terms of temperature. It would be really annoying if your seemingly stable overclock corrupts the video you are encoding or stops in the middle.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom