TPU's Core i7 Overclocking and Feedback

[Binge]

How does Cas 7 1600 mhz compare to Cas 6 1600mhz? I have not been able to get my OCZ cas 9 1866 kit to hit cas 6 yet, but I have not given them much voltage either, only about 1.68V, so 1.664V real and that works well for the cas 7 timings.

I don't know how the timings would compare. I only compared cas 8 1600 to cas 6 1600 to cas 8 2000.

I didn't think that voltage will really help timings, only frequency, that sound like it has any truth? lol. I dont remember where I read it from.

It depends heavily on which ICs your sticks have on them. If that were the case then 1066 cas 9 wouldn't tighten very well @ 1600 MHz, but I've run value ram at 7-8-7-20 before just by adding extra voltage.

 

[Chicken Patty]

hey guys, I just flashed to the latest BIOS of my board and never really tried to overclock with this new RAM I got. So i decided to try it out with the new BIOS.

This is where i'm at now. I know the vcore might be slightly higher than it should, i'm just trying to see how high I can go without having to raise vcore or if I do, not much more than that. So I set that as my max vcore desired.

This is my voltages in the BIOS, anything you guys notice odd, or that should be changed???

THis screenshot is under full load as BOINC is running in background. Idle voltage is 1.305v. 25 pass of Linx were ok.

I know timings are loose, but I don't want to mess with them till I get my CPU where I want it.

 

[mudkip]

Maby try to let LinX use more memory next time.

 

[mlee49]

CP, nice work but if your running your mem so loose why the 1.7V?

 

[Chicken Patty]

CP, nice work but if your running your mem so loose why the 1.7V?

its actually about 1.67v in windows.

you know what I just noticed, the timings are not supposed to be that loose. I think I need to set them manually. Also how much memory should I set Linx to use?

Question guys, what the hell is MCH Strap?

 

[t_ski]

Strap usually refers to what the NB is set at so it knows how to handle all the timings and dividers, etc. At least it used to on LGA775. If you set the strap at 400 you'd get a different series of timings and dividers (especially for the memory) than if you set the strap to 333, etc.

 

[Wile E]

Strap usually refers to what the NB is set at so it knows how to handle all the timings and dividers, etc. At least it used to on LGA775. If you set the strap at 400 you'd get a different series of timings and dividers (especially for the memory) than if you set the strap to 333, etc.

Yep. On most boards it was an automatic selection based on fsb speed, but on enthusiast boards, you could manually select the strap, regardless of fsb speed. So you could force the 266 fsb bus strap, on a 400fsb, if the board could handle it. (Lower strap setting gave tighter NB timings and latencies)

 

[Chicken Patty]

Strap usually refers to what the NB is set at so it knows how to handle all the timings and dividers, etc. At least it used to on LGA775. If you set the strap at 400 you'd get a different series of timings and dividers (especially for the memory) than if you set the strap to 333, etc.

so what setting is usually best for strap? anybody expirimented with it on the i7? I have it auto, maybe thats why my timings are so loose?

 

[mlee49]

IIRC the strap is the same as the DRAM ratio, so depending on what your strap is set to it may display the desired Mhz.

 

[4x4n]

New toy Seems to be a pretty decent chip. Did this yesterday when temps here were at 85f. I little more voltage than I'd like, droops to 1.34 at load. Running now at 4.0 with 1.25v. This would be a great chip for water, you can see I was hitting the 80's with my TRUE. Batch is 3903A472

 

[Baer]

New OC gadget and Bios for the REII

Hi everyone. It looks like a new BIOS 1501 for the RAmpage II E is about to come out, along with this OCing gadget.
Take a look http://www.theinquirer.net/inquirer/review/1433235/asus-rog-oc-station

 

[mudkip]

New toy Seems to be a pretty decent chip. Did this yesterday when temps here were at 85f. I little more voltage than I'd like, droops to 1.34 at load. Running now at 4.0 with 1.25v. This would be a great chip for water, you can see I was hitting the 80's with my TRUE. Batch is 3903A472

http://img.techpowerup.org/090712/21x200pass.jpg

Nice but how come 4,2Ghz needs much more voltage?

 

[n-ster]

Many D0s seem to need much more voltage after the 4ghz mark... At least that's what I have observed... But 4ghz and less, the D0 can reach with very low voltages

 

[Binge]

Many D0s seem to need much more voltage after the 4ghz mark... At least that's what I have observed... But 4ghz and less, the D0 can reach with very low voltages

You heard it from the expert guys I'm surprised that complex questions are being met with these strange answers. Isn't it enough that no two processors are the same? Even if two behave similarly they are rarely from the same batch, and for as long as i7 has been public we've seen such different results from everyone's chips. A lot of similar behavior like 4.0GHz easy, but nothing like 4.0GHz always under X voltage where X is some low number like 1.2V. Not all D0 are "low voltage" 4.0GHz. Not all C0 are worse than D0.

Most people here will forget any of this was ever mentioned. It seems like every time I find a way/an example to illustrate that these chips are not guaranteed low voltage, low heat, or bad OC, high heat, and then someone comes along and makes a simple observation, another someone comes along and makes a very simple minded generality based on the small # of chips he/she has taken the time to notice. Most of the posts in this thread are of good OCs, things people want to share to show pride. What some people forget is that our bed of chips is small, and some of us who have had worse experiences than others just continue to use what they have and keep humble. Rant rant rant... rabble rabble... stop assuming the best of the 920/Xeon D0s and wake up to the reality of things.

My C0 used 1.375V VCore and 1.26V QPI to get 4.2GHz stable, and my D0 does it at 1.24V VCore and 1.36V QPI. What changed? Nothing really. The heat output is the same. I was excited for a bit about it but now I'm rather non-plussed. I was lucky with both of these chips as I have seen worse.

 

[n-ster]

My C0 used 1.375V VCore and 1.26V QPI to get 4.2GHz stable, and my D0 does it at 1.24V VCore and 1.36V QPI. What changed? Nothing really. The heat output is the same. I was excited for a bit about it but now I'm rather non-plussed. I was lucky with both of these chips as I have seen worse.

Yes I have noticed this too... At the beginning, I was wondering if this could be just a big scam... like the same chips, but it SEEMS that it needs less voltage but in reality, 1.24V D0 = 1.375 C0 or something...

Everyone knows that all chips are different, but it is still noticeable that D0 GENERALLY seem to have an easier time getting to 4ghz on lower voltages... but D0s are sometimes overhyped indeed...

D0s can be worse than C0s, not all chips are the same.. No chips are EVER guaranteed something else than stock... Just that on average, some chips have either a good or bad chance of getting a lower voltage higher OCs etc... D0 aren't usually much, or any cooler than C0s... But more D0s use less voltages and get higher OCs than C0s... You can't deny that the D0 is still an improvement over the C0, but it is hard to say how much of an improvement it is.

You still have a good chance of getting a pretty good chip no matter what though... I mean, with a D0, I probably am sure most chips can do 4ghz on air without going say over 85 load... (Estimating a aoltage though is hard to do) Perhaps not all, but your chances are still high...

Whatever I just said can be completely false though, since I have seen very few OCs, and the i7s I did OC were quick OCs... Never Oced one more than 1ghz over stock...

 

[Assassin48]

that was a complete repeat of what Binge said lol

 

[Binge]

You're a good guy n-ster. I have always thought that. It's not about you or I being right, but it is about what is the truth. Props to your love for the i7. We all share it.

 

[n-ster]

Whoa thanks for the compliments lol Of couse we all share it And I think we pretty much all know your a great guy too

Yes it might just be a repeat lol..... That's good, that means we agree

 

[4x4n]

My 2 cents on i7 clocking.

I have had 5 chips, 2 C0 and 3 D0. I am using air cooling (TRUE and med speed panaflo), Gigabyte UD4P and 3x2gb OCZ DDR31600

All of them would do 4ghz fairly easily. The C0's take over 1.3v, the D0's around 1.25v. The C0's wouldn't go much past 4ghz. The D0's scale past that point but take quite a bit more voltage to do it. I've seen some posts in here of high clocks with lows volts, but I haven't been that lucky.

Also, I go 24/7 clocks. I could post some benches 4.5ghz, but that doesn't prove much to me. It's great if you're into just benching, just not for me.

 

[mudkip]

I can't sleep (5:45 AM here)

 

[n-ster]

It's not General Nonsense here

Is 3gb of RAM enough? The most RAM using thing I'll do is gaming on 1920x1080 if it makes a difference

 

[mudkip]

I have 3GB since the beginning I build this pc.

I never had a memory shortage 3GB is enough imo.

 

[Chicken Patty]

I have 3GB since the beginning I build this pc.

I never had a memory shortage 3GB is enough imo.

i couldn't play crysis with 3gb, major stutter. with crysis and other crap that windows used it took up about 2.85gb. that explains why it stuttered.

 

[mudkip]

i couldn't play crysis with 3gb, major stutter. with crysis and other crap that windows used it took up about 2.85gb. that explains why it stuttered.

I'm running windows 7 RC x64 with msn , wmp , firefox on and I've 70% of my memory left.

I know i can play crysis without stuttering

 

[Thrackan]

I found the first games that actually stutter when on 1680x1050 this weekend. Might be time to upgrade my 4850 after all...
Or, I could just increase my proc clock and see if it helps But I doubt that since AA/AF were the limiters.