EPoX "EP-9NPA+Ultra" NVIDIA nForce4 ULTRA Chipset Motherboard

[sixth]

gfnfgf - check your pm's....howd u get ur board to 2400mhz....i cant seem to get passed 2350 on mine with the x2, read your pm, i explained more there....thanks!

 

[sixth]

Anyone that is overclocking with this board with an X2 and hit 2400mhz on air hit me up, i need some help figuring out why my vcore is high to be able to be stable...thanks!!!

 

[imported_situman]

I could hit 2.75 on a Toledo core based 3800x2. If you have a Manchester core, you are pretty much out of luck pal. Toledo CPUZ should say E6, Manchester E4.

BTW anyone has the SLI version of this board and could share some experiences?

 

[sixth]

i have the mancherter, i know i have seen people hitting higher than 2.3 on a manchester....

 

[imported_situman]

The Manchester core I had sucked. I sold it rightaway.

 

[sixth]

damn it...this BLOWWWWWWS i needa freakin toledo!!! damn it! i can only get to 2350...either that or i am gonna say ****** it and jack it up to 2400 and run my vcore up to 1.65 and let her go....see if she burns up screw it....

 

[swamyg1]

So, would it be true to say that Manchester cores suck on this board? If so, WHY? Don't you guys think it might just be bad luck with either Manchester or Toledo? Why is this the only board affected by this when so many others get high OC's with either core? Is there a logical explanation, or is it all heresay?

 

[sixth]

Its hear say....how did you get to 2.45 on this board? I mean i am rock stable at 2350..but i gotta bump my vcore higher, what are you settings, such as vcore which is what i cant get on mine, where do you have yours set to? (exp + .125 etc)

 

[sixth]

This coming weekend when I get home I am going to try to change some things around, lower the muli to 9x and jack it to 270x9 which would give 2430 and hopefully lower my vcore, thats what my concern is...i think it might work.

 

[swamyg1]

I got my vcore at 1.475 on a 10 times Mulitplier and a 245fsb... HTT at 4X and I even upped the mobo voltage on the chipset to 1.6... don't know if that makes a difference or not. I think the main thing here is how your memory is matched with your cpu. I have a 166 divider which runs tight with my cpu clock speed. I messed around with different multipliers but found that this was the best config for me. It's funny, it was stable before (kinda) but it became rock stable when I lowered my CAS to 2.5 rather than 3... with the other mem settings intact. I also have my mem voltage at 2.9, I would like to get more out of this board (or cpu) but I don't really know what's holding me back, seems like I'm not the only one. I thought it was my xclio 450 but I dunno...

hope it helps

 

[sixth]

Wow...hmmmm i have mine pretty much the same but i think i have my HTT at 3x, chipset at 1.5 and my memory dividor at 133 (might be 166 not sure)...i havent changed my mem voltage at all......

But i cant figure this out becuase everytime i up my cpu voltage it goes to high and it gets to hot, i dunno what esle to do...

Last q....you say you have your vcore at 1.475, in the bios the only option there is adding voltage, like +.125 etc...what did you set yours to? This board is amazing, imean compared to my MSI this thing is great. I just think i hit a wall :-( either that or i am messing something up).

 

[swamyg1]

Ya my vcore is at 1.475... I believe the x2 comes stock at 1.45 or 1.5. I just upped mine +.25... I'm almost positive it's in the bios. Definitely up your mem voltage and try to push your mem a bit more... I'm sure that will help you.

I wish I could get a BETTER oc, it makes me jealous when I see guys hitting 2.6 to 2.75... I'm like WOAAAAAAH! It sounds like a bios fix would help us maybe.

So, are people having better luck with TOLEDOs or MANCHESTERs?!?!

 

[sixth]

Swamy...agreed! Hopefully a bios update will help pushing ours higher...

 

[phanna]

Okay, so I booted into memtest, and memtest says 227MHz as well. I am still not quite convinced as to what is going on here. I would like to know how these programs are deriving memory speed. Again, the POST screen says 245MHz. According to this article - http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2469&p=2 - the formula to calculate memory speed with Venice chips is DRAM Clock = CPU Clock / (ceil(CPU Clock Multiplier/Memory Divider)). So for me that would be DRAM = (2500 / ((8.5 / (5 / 6))) which comes out to 245MHz. So, either something is screwed up with my chipset hardware or something, or all of these programs are misreporting my memory. I am curious...can someone else with a 9NPA+ and a Venice chip test this for me? just set your divider to 333MHz (5:6), and boot into memtest or load CPU-Z, and tell me if it reports the proper memory speed. Thanks for the response

 

[teddyv]

Epox 9NPA+U arrived yesterday from the Egg and anxiously awaiting the case and a few other things from ClubIT. While I am waiting I thought I would download the drivers so I have the latest and greatest. A few questions...

1. What is the best Chipset Driver package to use? I know there is a new one out as of last week but primarily interested in most stable.

2. What Bios? Same interest as #1.

The Epox is being mated with a Venice 3000, MSI x800 128, and a Gig of Corsair value RAM. While we're at it, anyone have info on the latest ATi drivers (for the x800) as well?

Thanks folks!

PS - the drive cables are no longer hot pink thankfully, mine were white

 

[sixth]

I am running the latest of everything....Bios, october 24th dated i think, latest nvidia drivers, 6.70, and latest ati drivers...

 

[bwanaaa]

MY FSB (set in bios to 295) drops to 200 when I wake from standby according to cpu-z 1.30?!


EPOX 9NPA+Ultra 10/18/05 bios, AMD4800 x2, Ballistix 1T-3-3-3-7, XP Pro all the updates

FSB 295, memory 5/6 , multi 9x -> 2655 mHZ CPU, 245 mHZ memory

On boot, cpuz correctly reports the FSB and multi
sleep and then wake-> FSB goes to 200, multi goes to 10

Installing x2 driver does not change this behavior.

I have a feeling it's a flakey bios but this behavior persists despite reflashing the bios as well as reseting the cmos.

Is there a utility like clockgen that will set fsb AND multi? (as it is now, clockgen does not allow multi control)?

Obviously I would prefer an insight as to why this is happening so I can prevent it.

 

[sixth]

Well my machine NEVER goes into standby, and personally i think has something to do with overclokcing and going in standby, i dunno why i remmeber reading the somewhere.

 

[RobbyG]

Originally posted by: phanna
.....I am curious...can someone else with a 9NPA+ and a Venice chip test this for me? just set your divider to 333MHz (5:6), and boot into memtest or load CPU-Z, and tell me if it reports the proper memory speed. Thanks for the response

I'm running 245*10 with a 166 divider and everything shows as 204MHZ correctly for me (Post, CPUZ). I'm still using the 03/22/05 Bios if that matters.

 

[phanna]

I don't know. I'm still pretty sure its CPU-Z just misreading the divider. Just for the hell of it, I tried lowering my HTT frequency to 245MHz, and setting my divider to 400MHz (1:1). CPU-Z again erred, reporting a memory speed of 231MHz, which is CPU/9. It seems to assume that the dividers are based solely on the clock speed divided by an integer, rather than factoring the multiplier into the equation, which is the case with Venice CPUs and their on-die memory controller. Sometimes this may actually report correctly. When I boot at 9x200, CPU-Z reports a memory speed of 200MHz (CPU/9); this is the correct memory speed, but an incorrect way of deriving it. In the case of RobbyG, I suspect that CPU-Z reports your memory speed as CPU/12, (2450/12), which equals 204MHz. Using the equation supplied in http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2469&p=2 , the same result is reached, but I think that is just by chance. Either way, I am going to assume that this is a software issue, as my BIOS is correctly reporting my memory at 245MHz. Hopefully this issue will be resolved in future versions of CPU-Z.

 

[phanna]

Nevermind, I was ignoring one critical function in the formula. (ceil(CPU Clock Multiplier/Memory Divider)). The result of this portion of the formula is rounded up to the nearest integer. So, in my case, (ceil(8.5/(5/6))) = (ceil(10.2)) = 11. Therefore, it is CPU Clock / 11, which is 227MHz. CPU-Z is correctly reporting my memory speed. However, the POST screen is not. Maybe the updated BIOS fixes this?

 

[sixth]

maybe so...not sure...thats some nice math there haha i hate math .... but yeah than cpuz is doing everything right....

 

[raildogg]

hey guys, Im just about to order this board this second with Crucial Ballistix 2x512MB PC3200.

is this memory compatible with this board?

thanks

edit:

just checked Crucial's website and it is indeed compatible

cheers

 

[bwanaaa]

Well, i found something interesting-depending on which utility you use to query FSB, you get different answers.
CPU-z 1.30 and ClockGen for nVidia nForce4 Version 1.03 will report an FSB of 200 upon wake from standby. Not surprising since they are both made by the same coders.
But so does crystal cpuID.

However, other applications like Central Brain Identifier 7.70 , GCPUID 2.0 RC, WCPUID and even the properties of 'My Computer' report the correct cpu speed.

ALL applications correctly report the fsb upon RESTART. It is just that on wake from standby that the numbers change.

Does anyone know why these apps are reporting different FSB speeds upon wake from standby?

 

[teddyv]

Can anyone explain why Windows would insist on using a 1394 driver to connect to the Internet and not the NIC? Been at this for the last three hours and am about to just start all over again tomorrow morning. The build was easy, from the drivers on it was a nightmare.