Sign in with
Sign up | Sign in

Under Core i7's Hood: Comparing The C0 And D0 Steppings

Under Core i7's Hood: Comparing The C0 And D0 Steppings
By

Whether you're looking at the IT hardware business or automobile industry, products often receive bug fixes, improvements, or upgrades during their active life cycle. Cars get face-lifted once they’ve been on the market for a while, usually at that halfway point between launch and the introduction of a successor.

In terms of silicon devices, we’re talking about steppings, which often incorporate a number of different improvements. When Intel introduced its Core i7-975 Extreme, it started to replace the 965 Extreme's C0 stepping with the more advanced D0 silicon. We found some significant differences between the two.

New Steppings Don’t Mean More Performance

First of all, we’d like to make clear that modified processor steppings eliminate known bugs, while also improving the production process. Typically, this includes optimization on a transistor level, which may result in extended clock speed margins, lower voltage requirements for standard operation, or even the addition of features.

As a result, the latest processors normally show better energy efficiency, as they may require slightly less power than previous steppings to deliver the same performance results. However, stepping updates usually do not introduce any performance increase, as the processor vendor would rather have you pay for that.

New Stepping Mean More Possibilities

Other publications, such as AnandTech (Core i7-920 D0 Preview) and X-bit labs (D0 Processor Stepping in Action), already reported impressive results with D0 parts over the previous, first-generation Core i7 C0 stepping. Anand looked at a model 920 while X-bit labs tested a 965-Extreme.

We decided to do an apples-to-apples comparison with both Core i7 Extreme editions: the 3.2 GHz 965 and the 3.33 GHz 975. We set both to run at 3.33 GHz in order to compare performance and, surprisingly, we did find differences. Then we tried to overclock as far as possible on air cooling using a Zalman CPNS 10X cooler. Finally, we performed power comparison measurements at the highest clock speed, which both the C0-based 965 and the D0-based 975 processor would support reliably.

Ask a Category Expert

Create a new thread in the Reviews comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 46 comments.
This thread is closed for comments
Top Comments
  • 18 Hide
    doomtomb , August 21, 2009 6:45 AM
    So in the conclusion, newer steppings = better product

    Not surprised
Other Comments
  • 18 Hide
    doomtomb , August 21, 2009 6:45 AM
    So in the conclusion, newer steppings = better product

    Not surprised
  • 0 Hide
    anamaniac , August 21, 2009 7:43 AM
    doomtombSo in the conclusion, newer steppings = better productNot surprised


    Yeap.

    Hoping my 920 coming is a D0 though...
  • 7 Hide
    WINTERLORD , August 21, 2009 8:04 AM
    hmmm did i miss something, or is there temputer specs on hyere? would be great to see what the temps are both at 3.33 and overclocked at 4ghz please, please, please :D 
  • 9 Hide
    ewood , August 21, 2009 10:30 AM
    so the 975 was stable at slightly less voltage and reached less than 200mhz higher STABLE clock. seems like that could very easily be due to variances in manufacturing and not the stepping itself. there have been many 965s to reach 4.2 or 4.25 stable on air which would have been higher than the 975. then again toms could have got a bum D0... I woudn't say from this article that the D0 is clearly superior to C0, but that's just me and my stat classes talking
  • 3 Hide
    Anonymous , August 21, 2009 11:13 AM
    On page 5 (Overclocking Core i7-965 (C0): 4.0 GHz) in the last image the cpu-z vt is 1.8V, for 3.743 - 4.009 GHz identical with the CPU PLL. I don't think that is correct.
  • 3 Hide
    sublifer , August 21, 2009 11:30 AM
    Quote:
    the D0 processor overclocked up to 4.27 GHz, but wasn’t stable until we reduced to 4.14 GHz. The older Core i7-965 Extreme couldn’t even do this.
    You make it sound bad that the 965 couldn't eke out another 3.5% to tie the 975. Whoop-de-do That variance could be seen between different 975's or 965's. 3.5% is not significant at all. The voltage improvements looked nice under load but again, you'll probably see the same variance between chips of the same model. Show us 10% plus differences and then maybe we can call it news.
  • 1 Hide
    Anonymous , August 21, 2009 11:31 AM
    I've purchased recently i7 920 and found out, that overclocked to 3.33GHz works absolutely stable on 1.024V under full load (Seti@Home). Consumption vent down by nearly 100W and temperature by 15 degrees centigrade. Marvelous, isn't it?
  • -3 Hide
    sublifer , August 21, 2009 11:35 AM
    Addition:
    At 3.5% performance improvement per year we'll see a gain of 10% in about 3 years. If you think about it that way its utter crap. I can appreciate improved steppings but as we can see, its nothing to get excited over.
  • 6 Hide
    adbat , August 21, 2009 12:53 PM
    I think underclocking should be a standard test - I like power efficiency and I do wonder how that would affect performance/stability
  • -7 Hide
    ALANMAN , August 21, 2009 1:11 PM
    Would have been nice to see benchmarks at all speeds.

    Thanks for another half-assed piece, THG.
  • -8 Hide
    Pei-chen , August 21, 2009 1:53 PM
    Throw the PII 965 BE (overclocked to 4.0GHz) into the ring. I am tired of people saying how good the PII is compares with i7.
  • 0 Hide
    eyemaster , August 21, 2009 2:34 PM
    I've read that steppings fix known bugs in processors but I never asked a question on it.

    If there's a bug in the processor, wouldn't it cause the CPU to give bad data or execute improperly? Possibly crash? How can a bug not be a bug?
  • 0 Hide
    ta152h , August 21, 2009 3:41 PM
    This is a really interesting article. Probably the people saying the differences are insignificant have the C0 i7, and are feeling queasy about it. They need to rationalize it.

    I think it's really significant when you can get so many small improvements that add up to a big one.

    I do have one complaint though. I'd have liked to see the power at 4.0 GHz for each. That would have really helped people see difference between the processors, since the power difference would have been far greater than at 3.33 GHz.

    For anyone reading this saying it's not significant, would you even want a C0 now, or would you make sure you were getting a D0? Yes, I thought so.
  • 0 Hide
    GoOakland , August 21, 2009 4:04 PM
    I went to the ark link but I'm still confused. How can you tell if you are getting a C0 or a D0?
  • 0 Hide
    WINTERLORD , August 21, 2009 4:15 PM
    to oakland, to get a DO stepping. you must cross your fingers and look to the sky :D  really you can't order a DO stepping so it's the luck of the draw. there are a few places that will gaurantee your stepping version but iv never personaly ordered from the off sites. only newegg where i buy stuff at.
  • 0 Hide
    Kill@dor , August 21, 2009 4:15 PM
    Nice review. I thought the 975 have much higher performance value over 965. I still think the difference in performance will show when you overclock to 5GHz on both chips...i'd really like to see those numbers.
  • 0 Hide
    WINTERLORD , August 21, 2009 4:16 PM
    also power at 4ghz and temps please :) 
  • 0 Hide
    marraco , August 21, 2009 4:39 PM
    Does the i7 920 behaves equally?

    the 920 is far more popular. Almost any benchmark is done with it.
  • 0 Hide
    NuclearShadow , August 21, 2009 4:48 PM
    As a owner of a i7 920 C0 @3.6 ghz I really don't care much about the new stepping its not like the improvement is gigantic like a whole new CPU. If you order one and get a C0 instead of a D0 its nothing to rip your hair out about they are still awesome CPU's and great to OC.
Display more comments