3.4E "fun"

[cornelious0_0]

Well guys, I'm actually kind of curious as to what's going on right now with the 3.4E. I know that there's going to be SOME weird-ass solution to what's happening, but it's pretty weird.

HeavyH2O (whom I bought the 3.4E from) was able to run at 250fsb with decent room temps, and I'm sitting at about 20C room temp right now, so one would think I'd be good to go. At first I ended up getting a $hitload of air in my heatercore from turning the case on its side with a closed loop...and it butchered my temps.....but now, after having it redone it still doesn't seem quite right. I dunno, its just that the 3.4E at stock voltage is idling at 32C in windows, and I'm pretty sure that my board is even reading a little low. Beyond that, if I bump it up, and get the cpu running at 1.5v or so, even at stock clock it runs at about 42C in teh BIOS.....just doesn't seem right. Even though the temps are still quite nice for a Prescott of this speed, it just seems like my h2O should be "doing" more.

This exact same system with my 2.8C clocked up to 275fsb without any issues, so I know that it's nothing with the memory, and most likely nothing with the motherboard directly. What's confusing is that if I set the voltage above 1.6v or so and leave the clock at 3.4GHz, the comp refuses to POST half the time.....and if I set it to run at 1.5v (what HeavyH2O said he used to hit 250fsb) and turn it up to 250fsb, no POST again, and sometimes it'll POST and just not "make it" into windows.

As of now I have yet to come up with a solution for this...all I do know that switching to the latest BIOS made it so that i could benchmark at 235fsb at stock voltage, whereas my previous limit was 230fsb.

Another thing that's kinda weird is teh fact that if I set the voltage too much higher then 1.6v and I do manage to get into windows, CPUZ says its running at like 1.18v or something weird, and teh system locks up when I try and run any benchmarks.

I'm not giving up hope on this thing, because it scales amazingly in AM3 and 3DMark2001, and even with my XT-PE at stock clock at 240fsb I'm hitting 78k.....and with the XT-PE at 630/615 and the 3.4E at 240fsb I'm at just barely over 28k in 3DMark2001.

If anyone has any suggestions at all I'm open to hear them, as I want to get this resolved as soon as possible...as I've never given up hope on an overclock, and I've still got plans for this puppy.

 

[tsuehpsyde]

Damn man...sorry to hear about this, I was so pumped to see your results lol. I'm not as much into hardcore OC'ing like yourself, but I'll throw what I can at the situation.

The first thing I'd try is (as much as a pain in the ass I can imagine it would be) is throw ur 2.8C back in and check ur temps with that, to make sure it's not the water-loop acting up. If that's not the case, then I'm kinda at a loss....I mean, I knew the prescott's were really damn hot, but that hot??!>!? What were HeavyH20's idle and load temps?

Also, I wanna tip my hat off to you, because all this time I thought u were like 23 or something maybe older, but you're in fact my age (im 19, i'll be 20 in a few weeks) and I'm damn impressed with what you do, makes me wanna try more stuff than I probably should with my PC

All in all, I hope you can figure out the problem...I don't know anything about water cooling, and the idea still scares the $hit outta me, and I'll eventually grow a pair and try it out, but NOONE around where I live is this into computers...and my one friend who kinda is is happy to run at stock and thing's OCing is stupid and pointless

Eh, other than that, I'd call DFI and ask them...sounds like the board might not like the E as much compared to whatever Heavy's running, or something. Wish I could help more than that...really wanna see that Intel kick some ass. Hell, your 2.8C was owning my 3400+ A64 Makes me sad to think about it....hope my system can't hear me out here, might get mad I said that and stop working

 

[cornelious0_0]

Thx for the support bro, I really hope that I do come to some kind of a conclusion.....

I have let HeavyH2O know what's going on, and I hope to hear from him fairly soon. I know that my idle isn't TOO much higher, so at the moment I think my water cooling should be fine. The only reason I think it might have BEEN the watercooling at least acting up is because my idle at stock clock and voltage was about ten degrees higher then it is know.

I'll let you guys know if I come up with anything or hear anything.

 

[Alex41290]

maybe you're getting some kind of voltage droop of some kind? the only thing i can think of would be a voltage problem...sorry if i didn't read this, but did you flash the newest BIOS again maybe? other than that, i can't think of anything...if i do, i'll let you know

 

[cornelious0_0]

I have tried every BIOS for my board, with the exact same outcome.

I have come to the conclusion that it IS the voltage causing the issues. As it technically WOULD clock up to 250fsb if I could give it a tad more power.....but anything much over 1.55v in the BIOS and the system either won't POST, or it will, and the cpu will be running at 1.184v.....severly dropping my temps, but limiting me to playing with paint on my comp, or else risk locking it up with the extremely low voltage being applied.

Since I have found out that it's basically just the voltage causing this I have sent DFI a more clear email directly pertaining to the voltage issue with my 3.4E, and I've also now got a thread up on XS asking about the voltage issues, if anyone else has ever seen anything similar with any of THEIR prescott systems, or if anyone'd know what I could do.

As I've already said, I'm FAR from giving up, so any help is appreciate...as I still have plans for this setup, and I'm not going to let some stupid thing like this get in my way.

 

[34th_H.uGe]

Have you think about the problem of the powersupply.

Maybe the +12V supply is short of.

 

[cornelious0_0]

Have you think about the problem of the powersupply.

Maybe the +12V supply is short of.

1: Hehe, not quite....I'm running the OCZ 520W Powerstream in my sig....and I've got the 12v set at about 12.4v, and it NEVER dips lower then 12.36v under full load.

2: Off topic, but your sig is 6 lines over the limit of 10 that we're all given...just thought I'd letcha know before you got a PM from "someone" else.

 

[HeavyH20]

The forum was down for a little, but, let's figure this out.

Here are all my settings via my Abit IC7-Max 3 MB for the 3.4E:

Start out at 242 FSB and 1.5875 V. This gave me 4117 MHz. This was my rock solid clock. I could always get this, no matter what the temps, so I used is as my baseline. Keep in mind that the reported BIOS voltage and the real voltage are not one in the same. The BIOS setting is usually higher than reality. My idle was around 46C at this setting, but that is according to the Abit BIOS, so that figure is probably high. Contrast that to a 3.4C that idled around 39C at 3.9GHz, and there should be around a 7C delta between the Prescott at idle and the Northwood. All memory optimizations were turned off, if possible.

When I go to raise the clocks higher, I set the CPU voltage to 1.6125. Any higher than this, I found it just created heat, no extra clocks. Now, that I have this much juice into the core, I clock the system in Windows via ClockGen to find the highest workable clock. Normally, I was hitting around 247 FSB, and occasionally, 250 FSB.

The Abit IC7 Max3 I had also had four phase power. Most MB's have 3 phase. But, the power requirements are not that much, so the ASUS P4C 800E Deluxe I had also powered it just fine. I think your MB is fine, there is just some setting causing the issue. You don't have PAT enabled anywhere?

I have the OCZ as well, and I found that keeping the voltages spot on helped the overclock. So the the 12V rail has 12.04 on it, the 5V rail, 5.01 and the 3.3 rail, 3.30. If I played with these settings, I found I just introduced instability.

I would also drop your RAM back to 4:5 to verify the CPU overclocks, that way you will not have two variables to deal with.

 

[VO_Nuclear]

This is a perfect example of why you cannot speculate on how parts will peform, even if they are used.

I would tend to believe or say that it is your mobo holding you back. I have always reverred the Max3 to be the best 478 board, and you'll always get better results with it with skill.

Does your voltage droop or drop heavily under load? I know you said it won't post over 1.55v, but at voltages under will it stay steady?

 

[cornelious0_0]

HeavyH20, I'll try the voltages....leveling them out a bit more, but I doubt it'll help things out.

I have actually figured out, and gone a little more in depth with what happens when I try to apply certain CPU voltages. For the results below, the left number is what I apply in the BIOS, and the right number is what it actually runs at:

1.55v --- 1.5v
1.575v --- 1.15v
1.6v --- 1.16v
1.65v --- 1.21v
1.7v --- 1.26v
1.8v --- 1.34v
1.9v --- 1.42v
1.975v --- 1.48v

As you can see, no matter what I do, it seems that the board is incapable of applying a voltage to my Prescott of over 1.5v. I'm not quite sure if there could be something as simple as some BIOS setting or "fix" to get around this, but it's kinda irritating at this point. At any above voltage there is VERY little fluctuation going on, and when they're set at something, they're stable from there on in.

It struck me as kind of odd that I still had/have access to my full 1.975v in the BIOS, and the 3.4E simply "unlocked" some lower vcore settings to use.

The weird thing is that I know that the BOARD is capable of supplying much more then 1.5v, as I had my 2.8C running on 1.95v steadily.....and I see no reason why it would be any different trying to apply more then 1.5v to the Prescott.

There HAS to be some stupid little thing I can do to get this working, as its to ridiculous of a problem to not have some kind of solution. I'm gonna go ahead and send DFI another EMail with the above voltage results I found and see what they say, as it makes no sense why I could apply more then 1.5v to my Northwood, but not to the Prescott.

 

[HeavyH20]

That is interesting. I wonder if somehow the DFI looks at the default voltage of the CPU and does voltage offsets from that value. Or, it may not be supplying enough amperage to the Prescott, which would drop the effective voltage. The Abit did have 4 phase power, so it was a little better than the ASUS I had for that very reason. If you want to trade your DFI for the Abit, let me know. I can send it via UPS Monday for you to try for a bit ( pun intended). If it works, great, do the trade if you want to keep it. I was going to build a cheap kid PC with the Abit anyway, so I do not need any special 875 board for that. I would rather see it keep on clocking for which it was intended.

 

[cornelious0_0]

That is interesting. I wonder if somehow the DFI looks at the default voltage of the CPU and does voltage offsets from that value. Or, it may not be supplying enough amperage to the Prescott, which would drop the effective voltage. The Abit did have 4 phase power, so it was a little better than the ASUS I had for that very reason. If you want to trade your DFI for the Abit, let me know. I can send it via UPS Monday for you to try for a bit ( pun intended). If it works, great, do the trade if you want to keep it. I was going to build a cheap kid PC with the Abit anyway, so I do not need any special 875 board for that. I would rather see it keep on clocking for which it was intended.

I actually want to take you up on that bro.....as if the MAX3 does work out for me then I'll ge MORE then happy, cus I know it was working out for you. The main thing is going to be making sure that my ram works fine in in MAX3, as it's a little more "riddled" with memory compatibility weirdness then some other 478 boards.

However, if the MAX3 "takes" to my ram and works out with the 3.4E then I'll send you the DFI for sure.

You still have my shipping info from when you sent the 3.4E?

 

[HeavyH20]

Yes, I have your information. I will send it out Monday. You should have it by Thursday. I also have copper heatsinks in the OTES setup instead of the cheap pressure aluminum ones. Keeps the power temps in line. The Northbridge still has the fan on it. I did not watercool that. There is also a new rev1.18 BIOS for it. I did not download it, but supposedly, it can run the 3.4E in a compatibility mode at 14X instead of the 17X. That way, you can bump the FSB to some crazy values. This should fix you up. I have the exact same PSU, as well, so we know what it will do.

 

[cornelious0_0]

Yes, I have your information. I will send it out Monday. You should have it by Thursday. I also have copper heatsinks in the OTES setup instead of the cheap pressure aluminum ones. Keeps the power temps in line. The Northwood still has the fan on it. I did not watercool that. There is also a new rev1.18 BIOS for it. I did not download it, but supposedly, it can run the 3.4E in a compatibility mode at 14X instead of the 17X. That way, you can bump the FSB to some crazy values. This should fix you up. I have the exact same PSU, as well, so we know what it will do.

Sounds good bro, except with the speed that the 3.4E moved at, if it gets out in Mondays mail i could have it as early as wednesday, so we'll see.

Do you know/have you heard if you can select multipliers from 14x up to 17x, or if its just 14 and 17 you can use....because 14x would put me back down to my old 2.8C clock levels, so if I was gonnd do anything with teh multiplier it'd have to be at 15 or 16 so I could hit some higher fsb's, still clear 4GHz, and have my ram topping out at least a little more.

I look forward to hearing from you on Monday, and I'll be sure to let you (and everyone else) know how things go.

I guess it's perfectly understandable that I'm having issues, as it is a completely different mobo, so with your same mobo AND cpu that you had running, as long as the board doesn't have any issues with my ram (most ppl I'm seeing with my ram in a MAX3 are clocking it quite nicely actually) I should be good to go, and maybe start to explore that "extra headroom" that we believe may be laying unseen in the 3.4E.

EDIT: This is where you can really tell the difference between boards like mine (that is "Prescott ready") and those like the MAX3 for example, that are more then just comaptible, but fully capable of a tweaked out "E" setup. I guess I can't be mad at DFI though, as the board does exactly what it's sposed to do, and DOES support the Prescott line, but I'll sure as hell be glad/anctious to get the MAX3 installed.

 

[Jonsey]

You can choose only the 14x or 17x multi.Nothing in between.

 

[Spike23]

I have a 3.0E and i thought i'd ask the question in here
I overclocked to 3.3ghz and left the fsb:ram ratio to 1:1
is that fine?

 

[Jonsey]

I have a 3.0E and i thought i'd ask the question in here
I overclocked to 3.3ghz and left the fsb:ram ratio to 1:1
is that fine?

Is it stable? If it is, then I don't see why not.

 

[cornelious0_0]

Is it stable? If it is, then I don't see why not.

I agree, as long as you can run what you need to/want to run on your computer without any issues you're good.

I always prefer to do some "proper" stability testing with Prime95 with my overclocks, but not everyone is into this to the same degree.

I highly doubt I'll be using the 14x option with the MAX3, as I'll just be back to where I was with my 2.8C, with the extra cache basically.

I'm eagerly waiting for the MAX3 to get here now, as I want to at LEAST gain stability at 250fsb, but seeing as HeavyH2O never got to test the 3.4E at anything higher then that because of his ram who knows the fun I might be able to have.

 

[Spike23]

Is it stable? If it is, then I don't see why not.

Yea its stable, i got a suggestion from another person on the forum just to raise my vcore.
Cuz when i was using Prime95 says crtical error 0.5 off from 0.4.

It hasn't crashed yet. Only thing is the Pirime 95 makes me wonder
Is it my ram? Because i increased my voltage to it to like 1.70 and the prime 95
said criticla error 0.48XXXXX from 0.4

Hmmm should i get better ram or something? I even increased my cpu voltage but that didn;t help lower the 0.48XXXX when i increased my cpu voltage only made it go back to 5.0

 

[Jonsey]

Well, I finished water cooling my 3.4E this weekend as well. I had an AS8, which is sort of a socket 775 865 board. Anyways, the thing seemed stable at 3.9 in prime 95 (I didn't test it too long) but somehow I lost keyboard support. And then it wouldn't boot at all. Nice end to alot of work getting the watercooling up and going. I'll try again when I get a new board.

 

[eViL_M@LuM]

Sounds good bro, except with the speed that the 3.4E moved at, if it gets out in Mondays mail i could have it as early as wednesday, so we'll see.

Do you know/have you heard if you can select multipliers from 14x up to 17x, or if its just 14 and 17 you can use....because 14x would put me back down to my old 2.8C clock levels, so if I was gonnd do anything with teh multiplier it'd have to be at 15 or 16 so I could hit some higher fsb's, still clear 4GHz, and have my ram topping out at least a little more.

Your ram not capable of High FSB? I have a 2.8E, working on getting water setup. But it does 3.6 defaul voltages. I would much rather have 14 x 287FSB "my ram can do 292 stable @ 3.1V 3 4 4 8" Then 242 x 17. I'd figure as big of an overclocker as you are, you'd have DDR600 Ram by now .

 

[cornelious0_0]

Your ram not capable of High FSB? I have a 2.8E, working on getting water setup. But it does 3.6 defaul voltages. I would much rather have 14 x 287FSB "my ram can do 292 stable @ 3.1V 3 4 4 8" Then 242 x 17. I'd figure as big of an overclocker as you are, you'd have DDR600 Ram by now .

My ram is more then capable of what I'm throwing at it....as with a pidly 2.9v it hits 275MHz 1:1 with ease.

When I get the MAX3 I'll try maxing out the 3.4E with 17x and then I'll see how high I can hit at 14x also. In the end I wanna end up throwing a Vapochill LS on this 3.4E and going for my old 275fsb again, whihc would put me just short of 4.7GHz.....and considering I'll be hitting about 4.3GHz with the MAX3, I'd think that 4.7 should be quite doable with a modded MAX3 and the Vapo.

I am still waiting on a reply from DFI tech support about the voltage issue, as the board is what's causing the problem, end of story...and it simply won't supply more then 1.52v to my 3.4E. I dunno at this point if the board just technically can't handle it or if there's gonna be some kind of fix for this, but either way I'll be up and running again with my overclock before the end of the week.