A8N SLI/ 3200+ OC Help

[housecat]

I can go up to 265CPU freq with a low mulitplyer while running 1:1 on my current settings (2t, 2.5, auto on every other memory setting), with a 2.85mem volt.

When running 1:1 and a 10X multi, I can get to 250/251 without a problem. Then a single notch to 252 (or sometimes 251) and it crashes.
I tried bumping the mem voltage up, but it didnt help, and I tried the CPU voltage up to 1.625 and it didnt seem to help. Neither did 1.65v, that resulted in a lockup.

I've gotten the system to boot into windows before @2.55ghz.. so I dont *think* its the CPU, but I havent been able to duplicate that since. I think its the motherboard, as it makes no sense why my magical OC barrier is RIGHT at 250fsbx10.
Just seems to convienent for a CPU to max RIGHT THERE.

BTW I used 4x HT for ALL of this above (and all with stock ram timings). Even the 265fsb setting at the low multi. I figured above 270 (if I even reached it, which I didnt) would be where I'd drop it to 3xHT.


I think I could live with 2.5ghz though, I'll have to think if its worthy buying a new mobo and swapping this one out for 100-200 more Ghz. And I'm not sure the chip is even good for that.

I'd like your opinion on whats going wrong here at 251/252 and also if you think I should get a new board.

I really like the speed at 2.5ghz, as it feeds my SLI well... but I'm itching for more.



I suppose I could change it from 1:1 to a lower divider to see if taking the memory completely out of the equation changes things, and I'm able to take the CPU higher.. but then its still the motherboard in question and theres no easy way (barring a swap) to check that! And I think its the boards faults causing problems at 250+...

EDIT- Right now I'm at
4x HT, 250cpu freq, 1.55vcore, 2.85v dram, 10x multi, 1:1
if i push to 251 right now, it will reboot usually. 252 always does.

 

[mickles]

Try 3x?

 

[housecat]

Wouldnt that just put the HTT down lower, when the rig already showed its stuff up to 265cpu freq x5multi and 4xHT? 1060HTT effective

I should add, I'm on 1004 final bios. Duvie said to try to boost the motherboard chipset voltage, but I dont see that setting for my mobo in the bios.

I have the bad feeling I'm at teh CPUs end, but it just seems really strange for that to be a nice, even, 2.5ghz

 

[GuitarDaddy]

Try using a lower CPU multi and a mem divider.

I have the 3200+ and the A8N-SLI, and the best I can get with the 10x multi is 245 x 10.
I can't even boot at 250 x 10, but if I drop the multi to 9x I can get 289 x 9 = 2601mhz. Using this setting with the DDR333 divider gives me a ram speed of 236mhz @ 1T.

You should always run memory at 1T if possible. 2T is a huge performance hit, I benchmarked 230mhz 1T against 287mhz 2T and the 1T setting won all benches by a large margin. I also read a post where a guy tested 1T and 2T extensively and he found that to beat 200mhz 1T (stock) he had to get to 295mhz 2T

The limit this board will run 1T is 250mhz, so use settings that will keep your ram between 200mhz and 250mhz 1T, this will beat anything you can run 2T.

 

[housecat]

Ok I'm at 1T now. 250x10.

So is this pretty good then? I've seen alot of ppl get 2.6-2.8ghz on their winnys/A8N.
I dont understand why with a 10x multi you have the problem going above 245, but with a 9x you can go all the way to 289? Is it a board quirk?

 

[Fenuxx]

Your proc may just be out of headroom, sometimes they just do that

 

[housecat]

From all I've read on this board, I just have a hard time with 2.5ghz being the PERFECT, magical roof....

It woudl also corrospond perfectly with this boards maxs that guitardaddy said: 250x10@1T... how perfect!
I just dont believe it, but if I get better overall performance this way, than at 270x10@2T, then so be it.

1T did seem to boost my CSS frames.

 

[housecat]

What do you guys suggest? Get the DFI? I'm kind of at my wits end.

Ive gotten the "windows errors", not blue screens, but 'we detected a error in a program and want to send it to MS" error.. is there any way to stop that from happening on my current OC?

 

[dashiki]

it is your chip not the mobo 2.5 is normal for a 3200 oc anything more and you got lucky. what week is your chip?

Dont complain when you can push a cheaper chip to a 400+ dollar range chip oc'ing isnt an exact science anything over stock is good and 2t is NOT the way to go.

Do you really want to rma a mobo for 100mhz? you WON'T notice a difference between them.

 

[GuitarDaddy]

Originally posted by: housecat
Ok I'm at 1T now. 250x10.

So is this pretty good then? I've seen alot of ppl get 2.6-2.8ghz on their winnys/A8N.
I dont understand why with a 10x multi you have the problem going above 245, but with a 9x you can go all the way to 289? Is it a board quirk?

The difference is the onboard memory controller of the A64. When you are overclocking CPU and RAM 1:1 you are putting more stress on the chip and it tops out quicker.

It has been proven time and time again that running 1:1 on A64's doesn't improve performance much at all. And CPU speed makes a much greater difference. ALL the top A64 overclockers on this forum use a memory divider(look at the SuperPI Contest thread and look at the leaders sigs)

My settings of 289 x 9 = 2601mhz DDR333 ram=236mhz 2-3-3-5, beat 250x10 1:1 2.5-3-3-7 by a sizeable margin.

Athlon XP's show a nice increase using 1:1, A64's don't

 

[housecat]

Thanks guitardaddy, thats all good to know.

So I'm better off trying to see if I can get the chip to go higher than worry about staying 1:1.

But you said that running 1T is the way to go, and that this board wont go above 250@1T.. so I have maxed out my performance right?

Are you running 2T on your settings?

 

[Dough1397]

Originally posted by: housecat

I think I could live with 2.5ghz though, I'll have to think if its worthy buying a new mobo and swapping this one out for 100-200 more Ghz. And I'm not sure the chip is even good for that.

100 ghz amd 64 would murk everything.... but i doubt it will happen

 

[housecat]

lol


also, i had to put it to 1.6v to get 2.5 stable.
it runs 2.5@1.55 but i get errors.. i didnt mess with the intermintent settings i just went straight to 1.6v.

but cpuz shows the voltage varying quite a bit, i see it go from 1.6 to 1.584... it doesnt do this when i set it at 1.55v

 

[Duvie]

Originally posted by: housecat
Thanks guitardaddy, thats all good to know.

So I'm better off trying to see if I can get the chip to go higher than worry about staying 1:1.

But you said that running 1T is the way to go, and that this board wont go above 250@1T.. so I have maxed out my performance right?

Are you running 2T on your settings?

No matter if you run a 166 or even a 133 divider??? No matter on the vdimm??? No matter if you run cas 2, 2.5, or 3???

I find it hard to blieve 1t will just tail off like that......I am running 1t in my settings...I can'r un 1t stably at cas 3,4,4,10 at 1:1 at 296fsb though....It help[s if I drop to 2t but then I basiaclly have the bandwidth of cas 2.5 at 100mhz ddr slower so might as run it slower since as guitar daddy as pointed out...less stress on the memopry controller and likely better stability....

 

[housecat]

I'll try some of those suggestions like the divider. I guess I might not be grasping all the concepts here, but I dont see how running a divider will help things when my memory has shown good at these settings up to 530mhz. But I guess it has something to do with the motherboards defiecencies not my ram.

the markings on my chip were:

ADA3200DIK4B1
CBBID 0451FPPW
1173387A50112

I dont know what that all means, but i thinks a week 45?

 

[GuitarDaddy]

Yeah the 250mhz brick wall running 1T is the only thing that really bites on this A8N-SLI board. I'm starting to give up on a bios fix. But its not worth switching to another board IMO unless your a competive OC'r or something.

 

[Duvie]

Originally posted by: housecat
I'll try some of those suggestions like the divider. I guess I might not be grasping all the concepts here, but I dont see how running a divider will help things when my memory has shown good at these settings up to 530mhz. But I guess it has something to do with the motherboards defiecencies not my ram.

the markings on my chip were:

ADA3200DIK4B1
CBBID 0451FPPW
1173387A50112

I dont know what that all means, but i thinks a week 45?

My ram will run 8x283 1:1 cas 2.3,3,7 fully stable and memtested yet wont run at 9x283 even though the system is stable at default vcore....It wont do 1:1 at cas 3,4,4,10 at that level.....It can run 8x310 cas 3,4,4,8 stable yet wont d0 3,4,4,8 at 295 1:1 needs 2t and then may not be fully stable....

 

[GuitarDaddy]

Originally posted by: housecat
Thanks guitardaddy, thats all good to know.

So I'm better off trying to see if I can get the chip to go higher than worry about staying 1:1.

But you said that running 1T is the way to go, and that this board wont go above 250@1T.. so I have maxed out my performance right?

Are you running 2T on your settings?

I,m running 1T. Just use a divider and keep your memory less than 250mhz, and raise HTT as much as you can

 

[GuitarDaddy]

Your markings show its a week 51. 0451= 2004 week51

Many people say thats a bad week. But your OC numbers look just like my week 41 numbers and I've seen others getting great overclocks on 0451

 

[GuitarDaddy]

Originally posted by: housecat
lol


also, i had to put it to 1.6v to get 2.5 stable.
it runs 2.5@1.55 but i get errors.. i didnt mess with the intermintent settings i just went straight to 1.6v.

but cpuz shows the voltage varying quite a bit, i see it go from 1.6 to 1.584... it doesnt do this when i set it at 1.55v

Yeah I always say the limit is 250 because some people claim they can get it stable? The actual 1T limit on my board (prime and memtest stable) is 246mhz.

The voltage fluctuations are normal at the max setting, which is what you've got there, mine fluctuates from 1.57 to 1.58 and occasionally jumps to 1.6

 

[housecat]

I dropped the divider down to 266mhz, 2T, 3.0cl, and I cant seem to get above 252 still. Thats on 1.6v and 2.85mem.

I guess its a CPU limit.. I guess.

Well if I can get 2.5 completely stable I'll be happy. Its good to know I didnt get a "bad" week51...

sometimes i think half this stuff is wives tales. we're on the internet in 2005, but somehow it harkens back into the middle ages deep on a mountain top somewhere in a forgotten part of germany or soemthing.

 

[GuitarDaddy]

Yeah looks like the CPU limit. Not bad, alot of people get stuck a 2.2 or 2.3

Ye needs some eye of the dualcore newt with a dash of black forest strained silicon

 

[sxr7171]

Originally posted by: GuitarDaddy

Originally posted by: housecat
lol


also, i had to put it to 1.6v to get 2.5 stable.
it runs 2.5@1.55 but i get errors.. i didnt mess with the intermintent settings i just went straight to 1.6v.

but cpuz shows the voltage varying quite a bit, i see it go from 1.6 to 1.584... it doesnt do this when i set it at 1.55v

Yeah I always say the limit is 250 because some people claim they can get it stable? The actual 1T limit on my board (prime and memtest stable) is 246mhz.

The voltage fluctuations are normal at the max setting, which is what you've got there, mine fluctuates from 1.57 to 1.58 and occasionally jumps to 1.6

Do you mean 250 is the 1T with 1:1 Memory timing? Because I think it's easy to get it higher on 1T with a mem divider.

 

[GuitarDaddy]

I know you can push past 250 using a divider with the AUTO mem setting but it forces 2T.
I've tested at 325 x 8=2600 ddr333 ram 287 2T.

But in my testing its a hard limit at 245-250 1T regardless of divider. If someone is getting over 250 1T using a divider, I sure would like to know? Please post a linky

 

[housecat]

This is only my cpu until dual core comes out.. so Im not too bugged by my results.

2.5 works.. but it does exhibit errors and lockups.. I dont know how to fix it so I'm running 240x10 right now, 1:1, 1T. Thats still pretty fast.
If u have any tips how to get good stability, go ahead and pass it on.. cuz its very borderline.. it boots fine, runs fine, and give no errors but then the rig will lockup in games like css.

I think its possible to get it more stable, as i had it stable in the past.. but changed things around to play some more.



I know people wont want to believe this, but I'm almost sure our results will increase moving ot the DFI... everyone I've seen on OCP and forums got results like we are getting.. they moved to a DFI/MSI and all of the sudden the results seemingly magically got better. I dont know if I'm going to do this, but I'm still thinking about it.
I'd wager that even if I dont get over 2.5ghz with the DFI, I bet it makes it stable.