Is the i7 920 D0 the new Q6600 G0?

[aamsel]

You just further proved my point though....

Many who don't bother to read into the details buy whatever cheap parts they can find, throw it together and expect an easy 4.0GHz without doing any real work or tweaking.

If that is your point then I agree with you 100%.

I have done two totally simple overclocks in my life.

One was the dual CPU motherboard years ago that had (2) Celeron 300a's
that overclocked to 450 by moving a jumper or two.

The second is this E8600 I have running 4.1GHz at just over stock voltage, but
I still spent a week or so testing it and finding the sweet spot, which I probably
never even really found, since I am convinced this chip might do more. But, yes
even with such a simple overclock I spent time, testing, reading and posting.

I don't think every CPU is better than every one before it, and there have been
landmark chips, such as the 486-25, the Pentium-M Dothan, and Core 2 Duo.

I don't know if anything Nehalem (Bloomfield or Lynnfield) is in the category of being
a landmark chip?

To me, the oddity is that people think they have a failure with a 920 @ 3.4GHz, and
they think they have a success with a 920 @ 4.0GHz+ which is not a lot of difference
in terms of computing power.

To spend time tinkering with an overclock is, for me, the norm.

But that's exactly my point. people buy these completely expecting to hit a certain clock without the knowledge or hardware to back it up. The only clock you should expect is stock.

That I don't agree with. A standard consumer buying a Dell or HP Nehalem should only expect stock, because that is probably the only BIOS option. A s/w overclock is all that is going to work for these folks.

But someone buying a 920 @ 2.66GHz and a known overclocking motherboard and known proper RAM knows that the chip is going to do 3.0GHz without a hiccup, and that is an overclock, albeit a small one. I have not read of anyone with a recent 920 that would not do 3.4 at reasonable voltage. It is that 600Mhz or .6GHz from 3.4 to 4.0 that determines if you
have a decent overclocking 920, and again, 3.4GHz from a chip rated 2.66GHz is not a bad overclock.

You are correct in that the consumer is only guaranteed 2.66GHz, but it is not going to bluescreen at stock voltage running 2.67GHz either!

 

[Gillbot]

That I don't agree with. A standard consumer buying a Dell or HP Nehalem should only expect stock, because that is probably the only BIOS option. A s/w overclock is all that is going to work for these folks.

But someone buying a 920 @ 2.66GHz and a known overclocking motherboard and known proper RAM knows that the chip is going to do 3.0GHz without a hiccup, and that is an overclock, albeit a small one. I have not read of anyone with a recent 920 that would not do 3.4 at reasonable voltage. It is that 600Mhz or .6GHz from 3.4 to 4.0 that determines if you
have a decent overclocking 920, and again, 3.4GHz from a chip rated 2.66GHz is not a bad overclock.

You are correct in that the consumer is only guaranteed 2.66GHz, but it is not going to bluescreen at stock voltage running 2.67GHz either!

Agree to disagree here.

When you buy a rated part, any REALISTIC person expects it to work at STOCK, no more no less. Anything extra is just a bonus. If you buy a 2.66GHz chip expecting more, then that is your mistake.

 

[WhoBeDaPlaya]

The i7, in general, seems far more complicated to overclock than 775. The i7 has always had thermal issues.

It's not that bad really. Just isolate each element (uncore, BCLK, mem, CPU, various voltages, etc.) as best you can and test.
Long time / pain-in-the-butt, yes, but it's doable.

 

[Rubycon]

The only time a CPU has thermal issues is when the system builder does not adequately design a thermal solution to deal with increased TDP (in the case of overclocking at higher voltages, etc.)

This "problem" plagues ALL cpus not just the i7.

 

[WhoBeDaPlaya]

*Glances over at old SFF Dell GX270 in the corner*
Yup - 3.2GHz Northy in a tiny a*s case. Lots of bulging caps when I got it (replaced) and the HDD would damn near cook itself. Throwing in a few Delta 50mm fans solved the problem

 

[Gillbot]

Example....

 

[Painman]

I don't have an i7, but I suppose I bit into the hype somewhat when I got my Q6600 G0. Partly a lack of research too - I had a sudden "requisition order" come in from a family member for a new machine, my old A64 stuff actually filled their bill, so I took the opportunity to upgrade.

Ended up with a chip that'll do 3.2 comfortably, 200-300 MHz more if I cared to futz with it, but I've had this setup for a year now and haven't really felt the need to futz, let alone go with a 920 and play the lottery all over again. This machine is still quite fast enough for me, even at "only" 3.2.

 

[Grim281]

I have built plenty of oc'd rigs over the years and most chips "could" hit that magic 3.6 or nowadays 4.0ghz, but they usually require a massive jump in voltage to achieve it. My current 920 D0 system can do 3.8 with HT on at 1.225. For me to hit 4.0 I need to bump it over to 1.3 or above. The extra 200mhz is not worth the extra juice in my opinion. I stop my oc's when I reach that point.

 

[SanDiegoPC]

I have built plenty of oc'd rigs over the years and most chips "could" hit that magic 3.6 or nowadays 4.0ghz, but they usually require a massive jump in voltage to achieve it. My current 920 D0 system can do 3.8 with HT on at 1.225. For me to hit 4.0 I need to bump it over to 1.3 or above. The extra 200mhz is not worth the extra juice in my opinion. I stop my oc's when I reach that point.

I agree. My 920 is at 3.8 using only 1.2V (almost identical to above). Stable? Yea - it's been on for 10 days straight as of this morning. I haven't shut it off since restarting it with the new True CPU cooler and finalizing the overclock.

So it's certainly stable enough for me! And coming from an overclocked Q6600, I'm pleased as punch with this new build. There is no comparison between these two computers. At all.

 

[Gillbot]

I think there are too many variables and unreasonable expectations in the OC world honestly. Everyone just EXPECTS a D0 to hit 4+ with HT on at a low voltage with minimal cooling.

 

[ExcaliburMM]

I can say that, because I grabbed it! ():awe:

I got one of the open box ASRocks too. Due here Tuesday.

 

[rolodomo]

I have built plenty of oc'd rigs over the years and most chips "could" hit that magic 3.6 or nowadays 4.0ghz, but they usually require a massive jump in voltage to achieve it. My current 920 D0 system can do 3.8 with HT on at 1.225. For me to hit 4.0 I need to bump it over to 1.3 or above. The extra 200mhz is not worth the extra juice in my opinion. I stop my oc's when I reach that point.

Good plan and similar to what I experience with my D0. If you're not subject to cooling limitations however, why not pile on the voltage, but just make sure the O/C is stable using speedstep? It takes me 1.4V to reach 4.2 (4.0 HT on) with speedstep on and everything stable. So here I am writing this message and the voltage is at, let me check...1.184V (x12/2.66 GHz), which is very reasonable. That's where the voltage is most of the time until I actually need the speed.

 

[Thor86]

I think there are too many variables and unreasonable expectations in the OC world honestly. Everyone just EXPECTS a D0 to hit 4+ with HT on at a low voltage with minimal cooling.

Not sure where you are basing this information from? Is this just your assumption? If people think this, then they do not know what overclocking is fully about.

 

[Gillbot]

Not sure where you are basing this information from? Is this just your assumption? If people think this, then they do not know what overclocking is fully about.

Just take some time to read across various forums. Everyone is always asking how to get their 920 D0 to 4+ GHz.

 

[Thor86]

Just take some time to read across various forums. Everyone is always asking how to get their 920 D0 to 4+ GHz.

Asking and expecting are two different things, IMO. Irregardless, OCing is relative term (it's different with everyone) and never will be a guarantee, unless you go with some vendors that do with a premium pricing.

Is thread is debating on things that are too generalized. Again, IMO.

 

[jimhsu]

So the real question is: in air-cooled systems, is the i7 920 heat-limited or voltage-stability limited? My E8400 is definitely the latter - I start getting Prime95 errors even though temperatures are only around 70C or so. It's not a bad overclock by any means, but my chip refuses to do 4 Ghz stably for some reason (3.9 is fine, but it's hot in a SFF system). Then again I refuse to push my CPU to potentially damaging voltages, so perhaps it can go higher.

 

[Thor86]

So the real question is: in air-cooled systems, is the i7 920 heat-limited or voltage-stability limited? My E8400 is definitely the latter - I start getting Prime95 errors even though temperatures are only around 70C or so. It's not a bad overclock by any means, but my chip refuses to do 4 Ghz stably for some reason (3.9 is fine, but it's hot in a SFF system). Then again I refuse to push my CPU to potentially damaging voltages, so perhaps it can go higher.

Hmm, this is off topic, but anyhow, what kind of air flow does your SFF factor have? I have yet to see any SFF with good air flow, so any voltage increase will obviously be heat limiting in your case. Again, this is just my conjecture, and it very well could be the limit of your cpu.