• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

TPU's Core i7 Overclocking and Feedback

If its a latter D0 batch, I think Intel figured out how to sort them and keep 975EE chips ending up being branded as 920s, etc. It seems that the chances for better 920s are batches pre 39xx. My 3903 will do 21x200 at just under 1.3v, I think. I need better cooling to test as stressing to find stability at that voltage will make the chip run in the high, high 80s for most of the time.

what about the batch 3849A866 :D :toast:
 
that's a good point I dont know how INtel will determine what is an extreme edition chip and which isn't, I thought it was where they cut on the wafer and what passed the tests and was binned has a high performer.
 
Capture016.jpg

Low volts FTW
EDIT: Wow CP our batches our almost the same, mine is 3849A860
 
If its a latter D0 batch, I think Intel figured out how to sort them and keep 975EE chips ending up being branded as 920s, etc. It seems that the chances for better 920s are batches pre 39xx. My 3903 will do 21x200 at just under 1.3v, I think. I need better cooling to test as stressing to find stability at that voltage will make the chip run in the high, high 80s for most of the time.

I think there is some truth to this. The last five DO's that I've owned were nothing like the first 6 overclocking wise, and the last 6 were 39xx...
 
Jake, it is an interesting procedure they go through to make chips. We watched a video in one of my lower level engineering classes. The video was probably from the mid 90s at best, so while some of the processes to make them have changed, most has stayed the same. They were showing how they build a chip by building it up and creating the gates for current flow.

Now, how does Intel tell between a 920 and a 965 D0? I can only guess where it would happen. There is no physical difference between the two, i.e. one is not a disabled version of the other. When you have errors building the transistor stack, you could end up with a partial which could be sold at a lower cost. Kind of like how some Phenom IIs can be unlocked and others can't. The ones that can't are because part(s) of the circuit that didn't come out as they should have. Since we don't have that situation, we have to assume that when Intel went to building 1366 chips from the initial die, they were able to build 2 full and working chips initially able to become any one of 1366 chips, from Xeons to i7s.

From what I can remember from the video, there two tests after the chips get built and look as if they should work. Basically, you can build the chip and it look alright, but when you go to apply voltage, a transistor(s) doesn't work and causes the chip to not fall with in the allotted voltage specifications or not work at all. I can't remember what the two processes that the video showed, but the basic is this: one was to test for voltage while being in an oven and cooking, the other was to test to check instruction errors that they programmed the chip for in a similar environment.

I'm sure Intel has other final steps for deciding which 1366 gets labeled as a 920, 950, or 975EE. As this is a new architecture for them, and the first with an IMC, it doesn't really surprise me that it has taken this long to get it figured out. For all we know, the same thing might happen again when they go to the next stepping. Eventually, there will end up more consistency like that we have seen with the Core 2s.
 
I think there is some truth to this. The last five DO's that I've owned were nothing like the first 6 overclocking wise, and the last 6 were 39xx...

mines a 39xx as well. Still badass ocer though, nothing like a 38xx though.
 
Hmm... i7 920 D0 batch # 3910A313
Benches at 4.57Ghz@1.392 HT on.
24/7 runs 3.8Ghz@1.18 HT on.

Am i just lucky then?
 
No Jupiter, I would say we are in the decent end. I havn't tried 4.57ghz, but I can do 3.8 at a very similar 1.18125v. That's at 19x200. I was hoping for a good 20x200 showing, but only after throwing some delay at it, I am able to do it at 1.26875v. The strange part is that I can boot in to W7 at as little as 1.2375v. I can even let LinX run but it quickly shows errors. Though, I ran a full Vantage pass at that setting. It takes 1.26875v to pass LinX when left running for hours. This is always with my standard 25000 problem size.

Maybe I'm wrong, but in my mind, I see being able to do 4ghz at 1.2v or less as having the good ones. Though, seeing more and more of what others need for 4ghz with their 39xx, mines not looking so bad.
 
My i7 is batch #: 3913B007
It will be interesting to see what batch # I get with the i7 920 that I should get next week.
I am going to have fun comparing the two:)
 
D0 #3845B??/

3,8Ghz 1.13125v (bios)
 
I have a 920 DO #3849B028 runs @ 4.2Ghz @1.32volts 24/7 (may go lower)
Will bench @ 4.6Ghz (maybe higher...Gets to hot much higher.)
 
The 1156 bracket just came in the mail for my Megahalems. Tonight I start benching my new i7 860. :rockout:
 
Sweet! I might be picking up an i5 750 and EVGA P55 FTW on top of my i7 860 and GA-P55-UD4P :p
 
The 1156 bracket just came in the mail for my Megahalems. Tonight I start benching my new i7 860. :rockout:

nice, keep us posted dude :toast:
 
Haven't run this bench in a while... I may actually enjoy Crysis Warhead with full detail... awesome

Capture006062.jpg
 
Sweet! I might be picking up an i5 750 and EVGA P55 FTW on top of my i7 860 and GA-P55-UD4P :p

:rockout::rockout::rockout:

I have a ASUS P7P55D Deluxe and a i5 750 along with a i7 920 and a ASUS P6T Deluxe V2, I'm gonna stick with this combo for a while. I love ASUS. :D
 
OK, Megahalems is installed, and here's a quick run of OCCT. For a first run, I gotta say I'm pleased.

occt860firstrun.png
 
What's got ya so pleased paulie?
 
What's got ya so pleased paulie?

Well, considering I'm on air right now, and my last 4-5 i7 920's have needed 1.35v or higher for 4.0ghz, I think it's looking good for a first run. Don't rain on my parade, Binge. :laugh:
 
Last edited:
Well, considering I'm on air right now, and my last 4-5 i7 920's have needed 1.35v or higher for 4.0ghz, I think it's looking good for a first run. Don't rain on my parade, Binge. :laugh:

but thats all he does :p

Temps are good considering air cooling. Time to higher!
 
its hard to be a heckler on the intrawebz.

most people cant tell the difference between heckling, sarcasm, or someone being an asshole.

usually they jump to conclusions and choose the asshole option.
 
Back
Top