UnOfficial EVGA X58 Classified Owners Thread

[dejanh]

Forget my last reply. Spent the whole day today meticulously putting my system back together for nothing. The dreaded 7f is back. This time pressing the BIOS chip does nothing. This board is officially toast.

I have to say that this is the single biggest disappointment since I started building computers and overclocking 13 years ago. The board did not even last 1 year. So much for paying top dollar for top product.

Very unhappy

[Garrett]

When I try to boot my 950 @ 225 BCLK it hangs just when it's supposed to show the JMicron SATA/PATA stuff... do I just want too much or...

When I'm in Windows I can not OC the BCLK above 221 without the system hanging...

I'm on water, Heatkiller 3.0, PA120.4 and XSPC RX240 so cpu temps are fine, and the board has stock cooling but with an Antec Big Boy 200mm fan blowing on it with both side panels off...

Anyone else able to boot up with BCLK 225 or above with air/water?
I've tried S61A and S61B... kind of lost at the moment, could be a setting I overlooked though...

[zalbard]

You need high PCI-Express frequency, and also cold most likely.

[The_Game]

Anyone else able to boot up with BCLK 225 or above with air/water?
I've tried S61A and S61B... kind of lost at the moment, could be a setting I overlooked though...

If you are trying with high multi and 8 cores may be this is the problem. My limit is also arround 225bclk when i`m using 21 multi and 4 cores with HT enabled. When I disable 2 of the cores I am able to boot 21x230, also 12x240.
BTW I`m on water.

[Garrett]

Yes I think you're right The Game, but I did read that some people could only boot with 220 and push all the way up to 235 or even 240 in Windows...
I wouldn't mind bumping it up in Windows since I don't need to run more than 220 BCLK anyway for 24/7, just for benching now and then, but since it hangs at 221 I'm not happy yet about it...
I don't think watercooling the NB would help a lot here, right?

[The_Game]

I`m not cooling the NB its even without fan. The only thing i`m cooling is the digital PWM, but there is no big diffrence whether i`m at 5 ambient or 25 ambient the max bclk (~240 with 2 cores is always there) with 8 cores sometimes booting at 220, sometimes at 225(with 8 cores I`m always using 21 multi, never tried lower) but its because the CPU is cooler not the board, and yes i can reach more using eleet but when i run multithread bench it freezes or reboots which is normal when i cant even boot in windows successfully .

[Viss]

With bios S61A & B booting over 220mhz shouldnt be a problem anymore (before it was). PCI-E at 105~110 and CPU Impedance at "Less" should do the trick.

[Bojamijams]

Can anyone test something that I've witnessed with LinX thats quite interesting.

I'm stable at +100mv VTT in LinX. the GFLOPS is 45.

If I go to +125mv VTT, my GFLOPS increases to 49.

If I goto +150mv VTT, gflops goes to 52.

So I'm getting better floating point calculation by giving the VTT more voltage, even though for stability, I don't need it. It looks like it scales its performance based on VTT voltage.

I'm curious if anyone else is seeing the same things and if they see the performance transfer anywhere else?

[Garrett]

With bios S61A & B booting over 220mhz shouldnt be a problem anymore (before it was). PCI-E at 105~110 and CPU Impedance at "Less" should do the trick.

hmm interesting... is yours the 759 or the 760? (I guess it shouldn't matter right?)
ps you got messenger right?

[The_Game]

Yeah he is right, before S61A the board was not able to boot at more than 222bclk and the key was 220 from BIOS and rising with eleet, with S61A/B there is no problem I can boot 240 from BIOS . And yes its for E760, but there is equivalent BIOS for E759 too.

[bullydog]

Yeah he is right, before S61A the board was not able to boot at more than 222bclk and the key was 220 from BIOS and rising with eleet, with S61A/B there is no problem I can boot 240 from BIOS . And yes its for E760, but there is equivalent BIOS for E759 too.

240bclk on 4 cores HT enabled ?
What sorta vcore and PCI-E Frequency do you need to get it to boot at 240bclk ?

[The_Game]

240bclk on 4 cores HT enabled ?
What sorta vcore and PCI-E Frequency do you need to get it to boot at 240bclk ?

240 2 cores HT disabled . Never tried 4 cores + HT with lower multi, but I guess my max will be arrond 230 I can try sometimes. Voltages 1.57VTT and PCIe 115, Vcore is not important, because I`m using 12x multi, with 21 and 2 cores I`m able to boot 230, no cooling for more.

[jcool]

Can anyone test something that I've witnessed with LinX thats quite interesting.

I'm stable at +100mv VTT in LinX. the GFLOPS is 45.

If I go to +125mv VTT, my GFLOPS increases to 49.

If I goto +150mv VTT, gflops goes to 52.

So I'm getting better floating point calculation by giving the VTT more voltage, even though for stability, I don't need it. It looks like it scales its performance based on VTT voltage.

I'm curious if anyone else is seeing the same things and if they see the performance transfer anywhere else?

That's weird, but LinX Gflops rating is bugged with HT on anyways (general Linpack issue with i7's and HT though). Try with HT off, do you see the same results? I would compare, just haven't gotten round to setting up the classy yet.
LinX also buggs out with Gulftowns BTW... doesn't heat them up properly like prime does.

[mrcape]

Fellas, great thread. I have found linx stability with 4.2ghz / 12gb mem and great temps (water). My only questionable setting is cpu vtt/uncore at 1.43v DMM read. vDIMM is at 1.645v DMM read.

All other volts and temps are low.

Can anyone confirm or deny that 1.43v cpu vtt is fine for 24/7? If there are mixed feelings can someone direct me to some hard data besides the intel spec?

Tx again for the thread and replies.

[dejanh]

Just a word of caution guys re: EK full cover block...

I have managed to resolve my problem and it turns out that the board was just fine but it was in fact the EK full cover block for the Classified that was causing the problem. Based on the instructions no spacers are provided or needed around the SB which based on my findings now is incorrect. Not putting spacers around the SB can cause minor flexing of the board when you tighten the screws which can in turn cause contact problems and all sorts of no post scenarios that look like a genuine problem with the board. I found out that tightening the SB screws even 1/10 of a turn too much to cause very small board flexing can cause quite big problems that may or may not exhibit right away. In my scenario, it took over 4 months before my problem showed up but was remedied by simply moving the spacers intended for the NF200 area to the SB screws. In any case, be careful when tightening and generally do try to put spacers around the SB more than the NF200 area or in case of the non-NF200 versions of Classified board the narrow channel between the NB and SB.

[dejanh]

Fellas, great thread. I have found linx stability with 4.2ghz / 12gb mem and great temps (water). My only questionable setting is cpu vtt/uncore at 1.43v DMM read. vDIMM is at 1.645v DMM read.

All other volts and temps are low.

Can anyone confirm or deny that 1.43v cpu vtt is fine for 24/7? If there are mixed feelings can someone direct me to some hard data besides the intel spec?

Tx again for the thread and replies.

Sorry for double post, I did not want to mix the previous post with this one...

Yes, that VTT is fine. I have been running VTT over 1.5V now for about a year (come 21st of November) without any issues. Of course, YMMV.

[mrcape]

Thanks man, I had read that 1.6v is abs max, so that's great that you have real exp with it at 1.5v for a year.

Also, I have the EK full cover block not yet installed and will be aware of what you said about the standoffs.

That's the beauty of these threads it's like the ultimate live support.

Tx guys!

[Sam_oslo]

Can anyone test something that I've witnessed with LinX thats quite interesting.

I'm stable at +100mv VTT in LinX. the GFLOPS is 45.

If I go to +125mv VTT, my GFLOPS increases to 49.

If I goto +150mv VTT, gflops goes to 52.

So I'm getting better floating point calculation by giving the VTT more voltage, even though for stability, I don't need it. It looks like it scales its performance based on VTT voltage.

I'm curious if anyone else is seeing the same things and if they see the performance transfer anywhere else?

Interesting finds. If VVT is affecting the performance in the way you are discovering, then it is really bad news for my W3520. It runs stable with VVT=1.21/1.26v(BIOS/Load) at almost any x8 RAM speed and x18 Uncore.

I couldn't track the effect of VVT on CineBench 10 score yet, do you get different results on CineBench 10 (or other CPU/system-performance monitoring SW)?

[kitfit1]

Can anyone test something that I've witnessed with LinX thats quite interesting.

I'm stable at +100mv VTT in LinX. the GFLOPS is 45.

If I go to +125mv VTT, my GFLOPS increases to 49.

If I goto +150mv VTT, gflops goes to 52.

So I'm getting better floating point calculation by giving the VTT more voltage, even though for stability, I don't need it. It looks like it scales its performance based on VTT voltage.

I'm curious if anyone else is seeing the same things and if they see the performance transfer anywhere else?

I get the same on my rig as well. I also got the same results on the DFI UT x58.

I must say though, it has no effect at all on any other benchies. But it does have a very large effect on CPU temps when running Prime or Linx. So as far as i'm concerned, the lower the VTT the better. I remember reading a post on another forum where the OP was claiming 60+ GFLOPS at 4.2ghz. I found that odd, as my 24/7 clock is 4.3ghz and in Linx i get 58 GLOPS. I then realized he had done it by running VTT at 1.5v+. God knows what his cpu temps were running VTT that high for only 4.2ghz

[jcool]

Thanks man, I had read that 1.6v is abs max, so that's great that you have real exp with it at 1.5v for a year.

Also, I have the EK full cover block not yet installed and will be aware of what you said about the standoffs.

That's the beauty of these threads it's like the ultimate live support.

Tx guys!

I killed the 2nd Ram channel of a 940 C0 after 3 weeks 24/7 crunching at 1,53V VTT - on phase change
I wouldn't say 1,5V VTT 24/7 is safe. More like 1,45 absolute max, if you can keep it at or below 1,4V.

[dejanh]

I killed the 2nd Ram channel of a 940 C0 after 3 weeks 24/7 crunching at 1,53V VTT - on phase change
I wouldn't say 1,5V VTT 24/7 is safe. More like 1,45 absolute max, if you can keep it at or below 1,4V.

Well, I do not really agree, but that's my opinion. I like to run really high Uncore (don't ask why, it's more for s and giggles, not any particular reason) and I ran my VTT ~1.5V for a very, very long time on several chips. I did not have any problems. I think that it is a toss up and it will vary chip to chip, hence why I said YMMV. My opinion does not change though...from my experience high VTT is okay.

[mrcape]

Thanks guys all the data/findings are appreciated. I will certainly set it at the lowest value that allows me to remain stable at the same clocks when time permits. I def do not want to fry my memory controller, and the IMC thing seems uber fragile compared to separate MCH/NB.

[jcool]

Well, I do not really agree, but that's my opinion. I like to run really high Uncore (don't ask why, it's more for s and giggles, not any particular reason) and I ran my VTT ~1.5V for a very, very long time on several chips. I did not have any problems. I think that it is a toss up and it will vary chip to chip, hence why I said YMMV. My opinion does not change though...from my experience high VTT is okay.

Well what's there to agree, I just stated a fact (being that I had a CPU die on me due to high VTT). Like you said, some can handle it some can't. 1,5V VTT is definitely not a safe voltage for every chip, which is what you made it sound like.

[dejanh]

Well what's there to agree, I just stated a fact (being that I had a CPU die on me due to high VTT). Like you said, some can handle it some can't. 1,5V VTT is definitely not a safe voltage for every chip, which is what you made it sound like.

If I sounded deterministic I apologize. However, just because you had one failure also does not make higher VTT voltages unsafe. In fact, the whole VTT graveyard is not that rich. It could be that you just happen to have one chip that has a weaker IMC. So far I had 3 long-term chips that all comfortably did 1.5V+ on VTT without issues. If you want to be truly safe, then you are left with no choice but to stay within Intel's specs which also means 1333MHz DRAM max. I also never actually said that high VTT is perfectly safe. I made sure to use the word "fine" and add in "YMMV". Anyway...

To rephrase, based on my observations so far running 1.5V - 1.55V VTT for 24/7 use for longer periods of time is sufficiently safe so as to not warrant any reason for concern. It is however outside of Intel's specifications and thereby considered a risk.

[LardArse]

Just a word of caution guys re: EK full cover block...

I have managed to resolve my problem and it turns out that the board was just fine but it was in fact the EK full cover block for the Classified that was causing the problem. Based on the instructions no spacers are provided or needed around the SB which based on my findings now is incorrect. Not putting spacers around the SB can cause minor flexing of the board when you tighten the screws which can in turn cause contact problems and all sorts of no post scenarios that look like a genuine problem with the board. I found out that tightening the SB screws even 1/10 of a turn too much to cause very small board flexing can cause quite big problems that may or may not exhibit right away. In my scenario, it took over 4 months before my problem showed up but was remedied by simply moving the spacers intended for the NF200 area to the SB screws. In any case, be careful when tightening and generally do try to put spacers around the SB more than the NF200 area or in case of the non-NF200 versions of Classified board the narrow channel between the NB and SB.

oh, not just that, the ommitting the spacers can break the digital mosfets too, i've seen many boards with crushed mosfets and i know what was being done to these boards.