[S_Wilson]

I have had My 930 around 4 hours
SL94R
12/23/05 Pack Date
1.3V Max

Hardware used:

Asus P5WD2 Premium 955X, 0606 Bios
OCZ Powerstream 520 Power supply
2X1GB PQI DDR2 1:1 with FSB
MCW6002, MCP350 Pump, 2X MCR120 Rads - Stacked with 2X 120 Fans. Geared to be efficient and quiet and some say stupid, lol.


This chip is pretty much like Ross's, it goes well and then hits a wall where diminishing returns get pretty spectacular.

Using dual Prime95, large FFT as a guide here are some preliminary results.

3000 200X15 Volt=Auto Max Temp Per MBM5 = 44C Rad Delta = 12.1F

3750 250X15 Volt=Auto (1.3V) Max temp per MBM5 = 46C Rad Delta = +12.1F (25% OC 0% V Increase)

4005 267X15 Volt=1.4125V Max temp per MBM5 = 52C Rad Delta 17.1F (33.5% OC 7% V Increase)

4050 270X15 Volt=1.45V (35% OC 11.5% V Increase)

4200 280X15 Volt=1.55V (40% OC (19.2% V Increase)

It looks like the sweet spot is going to be somewhere around 4GHZ which is decent although 4.2G would have been better. It has been much easier to hit 4G+ than with any of 820's I have tried before.

I put the 270X15 in to show how dramatic the diminishing returns are, 2% more OC takes 4.5% more voltage!

I will update as I do more testing.

UPDATE
Further testing with dual Prime95 at 267X15 requires 1.425V with load temps going up 1C.

UPDATE 2
Things seem to be looking up, Have been priming at 4.1G, 273X15, 1.4625V in bios. Temps are relatively decent.

Best suicide SuperPI 1M run at 4.35G, 290X15, 1.55V in bios. 1 FSB more will not make it into Windows.

[hUMANbEATbOX]

1.55v for 4200mhz....that does not seem too good.

[S_Wilson]

Originally Posted by hUMANbEATbOX

1.55v for 4200mhz....that does not seem too good.

No, its not good.

[d94]

Originally Posted by S_Wilson

No, its not good.

[Ross]

Uh oh. I never even contemplated trying 1.55Vc, but that's probably what it would've taken for mine to be stable a 4.2GHz on air (4.0GHz @ 1.45Vc was way stable). Are you planning on staying on stock air, better air or do you have a WC setup?

I hope batboy and Ranger have better luck with the 940/920 they just got.

Good luck with it S_Wilson and keep us all posted

[dylskee]

Is 4.2GHz stable or are you not done testing yet? Well good luck to you S_Wilson, hope your results are better than we've seen so far.

[S_Wilson]

Originally Posted by Ross

Uh oh. I never even contemplated trying 1.55Vc, but that's probably what it would've taken for mine to be stable a 4.2GHz on air (4.0GHz @ 1.45Vc was way stable). Are you planning on staying on stock air, better air or do you have a WC setup?

I hope batboy and Ranger have better luck with the 940/920 they just got.

Good luck with it S_Wilson and keep us all posted

I was never on any kind of air, watercooled from the start although not aggressively.

I hope this is just a 930 phenomena and the 920 and 940 are good OC'ers.

[S_Wilson]

Originally Posted by dylskee

Is 4.2GHz stable or are you not done testing yet? Well good luck to you S_Wilson, hope your results are better than we've seen so far.

About 30 minutes of dual Prime95 Large FFT's. I am more interested in the 24/7 OC so I won't be exploring 1.55V any further, just wanted to see what it took and to illustrate the diminishing returns of this chip.

[Ross]

I hear that. It could also be this first round of Preslers, I'd really like to see a 950 since I am almost positive they'll be somewhat better OCers than any of the low ends, but I'd never spend that much on one anyway If the 920s/940s don't offer some ray of hope, that will clinch my decision for a 661 Cedar Mill instead

Good luck with yours...

[S_Wilson]

Originally Posted by Ross

I hear that. It could also be this first round of Preslers, I'd really like to see a 950 since I am almost positive they'll be somewhat better OCers than any of the low ends, but I'd never spend that much on one anyway If the 920s/940s don't offer some ray of hope, that will clinch my decision for a 661 Cedar Mill instead

Good luck with yours...

Thanks!

I want to see what the other lines will do also. I am a bottom feeder so there is no way I would spend more than I did to get a 930, lol. I am not too interested in the Cedar Mills, I prefer dual cores.

Overall, I am pretty satisfied. My goal coming into this was an easy 4.2Ghz 24/7 OC so I am not too far off at 4Ghz which is much better than my previous 820 attempts. Would have been nice to get a golden chip for once but I take what I can get.

[MikeyLikesItSI]

Wilson, where do you stand right now? or are you still searching for a 24/7 sweet spot?

[S_Wilson]

Originally Posted by MikeyLikesItSI

Wilson, where do you stand right now? or are you still searching for a 24/7 sweet spot?

Right now I am at 267X15 4Ghz, 1.425V 53-54C. It has been priming for over 2 hours so I have the feeling this will be the sweet spot.

In comparison, that is a little less voltage but about the same temp as my 640 @ 4.25Ghz.

I will probably let it prime a while longer then go for suicide runs with SuperPI and see how high it goes.

[RangerXLT8]

Originally Posted by S_Wilson

About 30 minutes of dual Prime95 Large FFT's. I am more interested in the 24/7 OC so I won't be exploring 1.55V any further, just wanted to see what it took and to illustrate the diminishing returns of this chip.

1.55v is alot man. That's dangerous IMO. I wouldn't even run a prescottt at that level for long, let alone 65nm.

[Okagi]

Originally Posted by RangerXLT8

1.55v is alot man. That's dangerous IMO. I wouldn't even run a prescottt at that level for long, let alone 65nm.

I used to run my Prescott at 1.55 actual constant. No problems even after months of use. I can't comment on the 65nm part, though. Take this for what it's worth.

Good luck guys.

[Ross]

Prescotts are also .10V higher default Vc. 1.55Vc would be the equivalent of running 1.65Vc on a Prescott, which is up there. 1.55Vc might be OK for a suicide shot since no one has seemed to have smoked any Preslers yet, but I certainly wouldn't plan on using it 24/7. I am such a wuss, I wouldn't even go over 1.50Vc on phase

[S_Wilson]

Originally Posted by RangerXLT8

1.55v is alot man. That's dangerous IMO. I wouldn't even run a prescottt at that level for long, let alone 65nm.

I think it should be ok for short testing runs with watercooling. I just wanted to see how much voltage it took to stabilize 4.2Ghz for 30 minutes of dual Prime95.

[S_Wilson]

Originally Posted by Ross

Prescotts are also .10V higher default Vc. 1.55Vc would be the equivalent of running 1.65Vc on a Prescott, which is up there. 1.55Vc might be OK for a suicide shot since no one has seemed to have smoked any Preslers yet, but I certainly wouldn't plan on using it 24/7. I am such a wuss, I wouldn't even go over 1.50Vc on phase

I am a wuss also for 24/7 use. I think 1.425V which is just over 10% increase is about as much as I want to use.

[Ross]

Yeah, I think 1.45Vc or less will be safe enough for 24/7 OCs.

[S_Wilson]

Well, the system has dual primed for more than 4 hours without issue 267X15 1.425V in bios. The CPU has a good feel to it at this level, it seems relaxed, not tense and I like it that way.

SuperPI with this config runs about 32.625s

I will continue tinkering and updating.

[S_Wilson]

Once the Colorado Avalanche were done embarrasing the Blues I have been able to get back to the matter at hand.

I am priming at 4.1Ghz, 273X15, 1.4625V in bios. The system still feels quite relaxed.

My best suicide run was at 4.35Ghz 1.55V in bios. I updated the original post with the screenshot. One FSB more and the system won't make it all the way into Windows, I get a stop error BSOD.