XP2500+ hit its ceiling?

[GHoosdum]

I finally got a decent fan to put on my Primary's SLK-900A, a ~60 CFM Enermax 92MM unit. Even with 60 CFM though, this processor is only doing 2.2GHz (200x11) at 1.875 VCore and running at 48C. That's too much for me, and it hard-locks at 200x11.5 - any suggestions?

[panzerkw]

some info on da memory plz

[Thrax]

I wouldn't suspect of the memory in this scenario by the simple fact that the chip is requiring almost 1.9v to do 2.2GHz.. Which is pretty atrocious for a Barton, but not entirely impossible.

Do you have the stepping of the chip?

[GnomeWizardd]

What bios version are you using?

[GHoosdum]

I've got the stepping in a text file somewheres on one of my PCs... I'll post it when I get home.

The memory is not a factor, except that in order to run synch I can't go more than 200 FSB because it's PC2700 RAM, But when I first OC'd the Barton I was running the RAM at stock PC2700 speed.

Funny thing is, this same chip was stable at 2.2GHz at 1.825v on the old DFI board I had, with the same PSU. Still, I'm rather disappointed in this chip since my XP2100+ cranks away all day at 2.3 GHz at a cooler temp at lower VCore and 10 less CFM on an SLK800A...

[Al_Capown]

That's a weak 2500, my old 2500 did 200x11 @ 1.65vcore (stock). No problems prime95 stable.

[keto]

Tried bumping fsb over 200 @ 11 multiplier?

Tried re-seating the SLk900?

[GHoosdum]

I'll try reseating the SLK900, and I can try to up the FSB, but I don't plan on keeping it that way, since I can't afford new/better RAM, and the timings are loose enough at 200 as it is, I don't want to drop them below the current 7/3/3/2.5

Yes, it's a weak 2500.

I'm using whatever version of BIOS shipped with the mobo... will a newer BIOS affect the OCability though?

[Al_Capown]

GHoosdum had this to say
I'll try reseating the SLK900, and I can try to up the FSB, but I don't plan on keeping it that way, since I can't afford new/better RAM, and the timings are loose enough at 200 as it is, I don't want to drop them below the current 7/3/3/2.5

Yes, it's a weak 2500.

I'm using whatever version of BIOS shipped with the mobo... will a newer BIOS affect the OCability though?

Yes a new bios will. Also change the timings to 11/3/3/2.5 should help with bandwith.

[GHoosdum]

I thought that lower #s were better in mem timings tho? When I try to set it to turbo mode (which the RAM can't do at 2.7V) it makes the #s 6/2/2/2...

I'll update the BIOS and see what that does for me.

[Thrax]

For the nForce2, 11/3/3/2.5 is better than 1-10/3/3/2.5

Optimal is 2/2/2/11.

[GHoosdum]

Well! I learn something new every day... I can't wait to see how that increases my PCMark score!

bwahahaha!

[pseudonym]

Thrax had this to say
For the nForce2, 11/3/3/2.5 is better than 1-10/3/3/2.5

Optimal is 2/2/2/11.

So I shouldn't be trying to set that thing at 6??? Hmmm, I'm with Ghoosdum here. Should I leave mine though if I'm doing 2/3/2/6??

[Thrax]

x/x/x/11.

Period.

Fill in the Xs with whatever else your RAM is capable of doing. Be it 2/3/2, 2/2/2, 3/3/3, 2.5/3/3... Just end in 11.

[ginipig]

Why 11, Thrax?

Will it work on any speed DDR? I'm at 2-2-2-5 @ 400, because my Cpu is running at 200x11. I've finally decided to try to get a little more out of 2500+a7n8x, since it's been quite stable these past few weeks.

Lets say I were to reach 217x _ , should I stick with Corsair's recommended 2-3-3-7?

[keto]

2-2-2-11 will always be the fastest on nforce2 platform if your memory will handle it. It's only a *tiny* bit faster than -5 or -6 but -11 is indeed faster. It's been tested a kazillion times all over the internet, Thrax has verified it on these forums, a number of other members including myself have verified it on our systems.

[Thrax]

No one knows why x/x/x/11 is faster than x/x/x/1-10, but it is. I chalk it up to the way the nForce2 memory controller handles its cycles.. 11 must just strike a chord.

This works on any DDR module in existence.

[ginipig]

I just needed 1 reply as an impetus.

2 just sealed it.

Going to test it now :)

[GnomeWizardd]

Def flash to a newer bios! I know i am using 1.8 and my old 2500 got too 220 FSB

[Disvengeance]

A great majority of bartons, like mine, hit 11X200 easily at stock voltages. Must be you have a bad chip or something. Although i would think even a bad chip would do better than that:(

[GHoosdum]

Here's the stepping, let me know what you think of it:
axda2500dkv4d
aquca 0323tpmw
And it's #231 in the batch :(

[GnomeWizardd]

AND I REPEAT! Update teh Bios!

[Geeky1]

Never heard of that stepping. Flash the bios (you can do it in windows now, you know, so you don't have any excuse :D) and see if that helps.

If not, I'd say that you just have a really rotten chip.

[GHoosdum]

I'll update the BIOS as soon as I switch over to that PC - I'm using my secondary right now and I can't put 'em on the KVM switch because my LCD monitor has DVI input...

Since my primary has no floppy drive, would you guys recommend using Abit's GUI BIOS tool for the update?

[Geeky1]

I'd just use FlashMenu or whatever its called, yes. I've used it before without a problem.

[Thrax]

Yeah. Almost everyone with an NF7-S here has used FlashMenu without a problem.

I swear by it.

[Geeky1]

I've used it before without a problem.

Almost everyone with an NF7-S here has used FlashMenu without a problem.


Now, of course, the next person that uses it is going to fry their BIOS...

[Thrax]

You win some; you lose some.

[GHoosdum]

I guess the next person would be me... :doh:

[Al_Capown]

It's pretty easy. Just make sure you don't have anything running in the background and click "one click live update" or something along the lines of that. I've used it with no problems at all.

[Thrax]

Searching around Google, I've found several threads on other forums that relate AQUCA Bartons made earlier than yours, GH, that are doing 2.4 and 2.5GHz.

Maybe it's time to look elsewhere; but I wouldn't get worked up over finding another culprit, you might have just gotten a worse-off one. :(

[GHoosdum]

Hmm... Hopefully the BIOS update will allow my chip to live up to its potential...

[GHoosdum]

OK - so I updated the BIOS. Fiddled around a little bit. Now the PC POSTs but won't go into Windows @ stock. It won't make it all the way into Windows at 11x133 1.75 VCore either, so I think something is wrongo here... suggestions?

PS - passed memtest. :p

[TheLostSwede]

It's all about cooling. I have learned that the last few days. My Barton wouldn't even run stable at 2.4 and 1.95 volt in win XP on Watercooling. With this Prometeia, same chip at same voltages primes fine at over 2650 (currently priming at 2693) in the same OS.

[GHoosdum]

So you're saying that an SLK900A with a 64CFM fan is insufficient to run it stock? ;)

[Geeky1]

Yes :D

[TheLostSwede]

I finally got a decent fan to put on my Primary's SLK-900A, a ~60 CFM Enermax 92MM unit. Even with 60 CFM though, this processor is only doing 2.2GHz (200x11) at 1.875 VCore and running at 48C. That's too much for me, and it hard-locks at 200x11.5 - any suggestions? Stock? ;)


So you're saying that an SLK900A with a 64CFM fan is insufficient to run it stock? ;)

IMO, the Thermalrights are overrated A LOT. They produce false temperatures by blowing air around the socket since approx. 40% of the effective airflow from a 92mm fan on that heatsink is going outside the heatsink. I know a lot of you guys will jump at me now but that's my opinion.

[GHoosdum]

Yup, stock:

OK - so I updated the BIOS. Fiddled around a little bit. Now the PC POSTs but won't go into Windows @ stock. It won't make it all the way into Windows at 11x133 1.75 VCore either, so I think something is wrongo here... suggestions?

PS - passed memtest. :p

[TheLostSwede]

Ok, didn't see that (not looking enough as usual). In that case, this is all of topic then. It's down to a direct hardware fault or problem here.

Before you flashed the bios, did you have Apic enabled or disabled? It's enabled by default. Windows will go tits up if it suddenly finds out that you are using Apic when the OS is installed with Apic off.

[GHoosdum]

Apic was on both before and after the BIOS flash. I'm rebuilding the PC from scratch in a new box, and clearing CMOS by jumper before trying again. I'm doing this tonight and I'll post my results, hopefully from that PC...

[MJO]

Yeah. Almost everyone with an NF7-S here has used FlashMenu without a problem.
I haven't used it yet.
I do not like the idea of flashing the BIOS while running windows.
There is a first for everything though. :werr:

Mackanz, why do you have to come here and say that the Thermalright is no good? :confused:
I am about to purchase one very soon.
A lot of people have had good results with it.
The Thermalright is better than a lot of other heatsinks right?

[GHoosdum]

I don't think he said Thermalright is no good - I think he said that a Prometia setup is better! ;D

Thermalright is a lot better than most other aircooling options though.

[MJO]

True, Thermalright is better than most.

It should be better than my ancient Silent Viking though.
The Viking is holding up though, I did not think it would suffice at my current speed.
But 2.2 Ghz - 2.4 Ghz cannot be acheived with it.
It is impossible.

BTW I moderate my comment.
I belive Mack said that TR is not as good as "everyone" thinks it is.
There that should do the trick.

And of course a Prometia setup is better *sigh" :doh:
If I had the dough I might go water as well.
But air is the cheaper way.

How come that two members are having problems with their NF7-S at the same time?
That is quite odd.

[TheLostSwede]

NF7 have about the highest percentage of rma rate at the moment of all NF2 boards actually. I don't have any numbers, but looking at 4 different forums is enough.

I'm not comparing with a Prometeia. That's just silly. But, one of these days i will prove that the TR thing is somewhat a myth. It's good, no doubt, PERHAPS one of the best available. But not as good as the reviewers says IMO. Put a 80mm fan at X CFM on that 900. Run 100% load for a week to give the goop some time. Measure the ambient temp at the same time. Then do the same with a 92mm fan at the same CFM.

IMO, the temps should be very close. I hope i am wrong though.

[GHoosdum]

I actually saw very close temps with an 80MM fan that is ~10CFM less than the 92MM fan that I use on it. In fact, I'm considering switching back from the 64CFM 92MM Enermax to the 53CFM 80MM Mechatronics, just to get more of the airflow in the fins.

[MJO]

Looking forward to see how you would kill that myth.

Regarding the NF7 rma situation.
Maybe Abit is putting all of their effort into the AN7 and their A64 boards?
It seems odd that the NF7 has become so bad in a matter of months.
2 months ago everyone was praising the NF7.
Looks like the tables have turned.

I am very happy that I have a NF7-S rev. 2.0 that is working flawlessly.

[Thrax]

I have 5 NF7-S boards, as old as April... And as new as this month. All of them work wonderfully.

[TheLostSwede]

The problem is that SOME of the new NF7 overclocks like beasts. MANY goes to the rma also. It's more luck than ever with this board. Nforce 2 chipset is the weirdest evar though.

[MJO]

My board is from April as well.
It is an early rev. 2.0.

Dunno how high the FSB will go yet.

I would prefer nForce2 over KT400A.
But the nForce2 sure has its quirks.
Here is a few of them.

Best memory performance with Tras at 11.
Performance drop when running async.

Anymore oddities?

It kinda reminds me of my old KT7A-Raid it was very fussy as well.
All the settings had to be right.

[GHoosdum]

Well, I'm posting from that PC now. It works, stock for now. The temp is a little high for my liking - 40C at stock voltage and speed. I'm going to try for 2.2GHz now and see what happens.

//edit: OC'd to 2.2GHz - does it at 1.8 VCore, but no less. It's an improvement, I suppose.

[Thrax]

Performance going to hell on an asynch bus is true of any platform for the Athlon.

[ngfrazier]

Hi GHoosdum,

I too am wondering if I have either a lousy Proc or Mobo? Please see my post here: http://www.short-media.com/forum/showthread.php?p=90149#post90149

While I do not have any problems running at stock voltage, I am depressed that I cannot take my 2500+ up to 3200+ stable (even with the voltage boost).

My stepping is the following
AMD Athlon 2500+ (retail box) with the following codes:
AXDA2500DKV4D KO22244L31038
AQZEA0332MPMW

I guess the big question is, "How do we narrow down the culprit once we know the RAM is OK: Motherboard, CPU, or Other"? Is the job of the Mobo in OC'ing to merely supply a constant stream of juice to the CPU, or can other components (like a faulty Northbridge) cause lousy OC'ing? Hmm...