What gives - NF7-S rev2.0 + mobile 2400+ = No 10.X multipliers? **Update**

[Jeff7]

Update:
I came across the XtremeSystems Forums while looking for power supply modding info (which it turns out I don't need - a multimeter made a liar out of my motherboard), and I found their Mantaray BIOSes.
I'm now running the D26 Mantaray BIOS. It properly identifies my Mobile CPU, and all multipliers are available now. I'd put my thanks there too, but I can't post anything there, despite having registered and activated my account.
So, I can try to run this mobile 2400 (1.45V, 45W) at 2.1GHz. It boots and runs up to 2.5GHz, but it won't stay stable for long. Once it's stressed, it goes up to 58C, even with an Alpha 8045, and then Prime95 chokes up a rounding error. Maybe it just can't handle the higher voltages (1.8v) it needs to boot at those speeds. I had it at 2.2GHz for awhile, which was at 11x, avoiding the 10.X multipliers, but it started crashing on me, so I guess it can't take that speed.
Or it might be the wildly varying voltages of the motherboard. MBM5 reports voltages ranging from 1.7V - 1.81V for the core.
But it's also inaccurate on the other voltages. MBM5 reports:
3.3V: 3.15 - 3.23V
+5V: 4.84 - 4.87V
+12V: 11.49 - 11.55V

The multimeter on the ATX power connector, while the system was running reported:
3.3V: 3.31V
+5V: 4.98
+12V: 11.99V

I wish I could measure what the CPU and DIMM voltages actually are at.
I wonder also why motherboards aren't made with better monitoring hardware. I mean, they've got incredibly advanced technology all over the place, but they can't do what any cheap voltmeter can do?



I found this a bit odd. I just got this CPU, and it works just fine at stock speeds (1.67GHz), and all the way up to 1.9GHz, BTW, where it's almost 2x faster at SETI than a 1.7GHz Tbred did, at least according to SetiSpy's MFLOP/s rating.
Anyway, as soon as I did 10x200, it wouldn't POST.
"Ok, maybe I got a dud that can't do 2GHz."
But then I tried 215FSB x 9.5 = 2.04GHz. And it POSTed. Granted, Windows barfed during bootup, probably thanks to my Value RAM. Still, it POSTed past 2GHz. So why not 10x200? 10.5 x 200 gave the same problem.

But now, it's running at 11x200. What's up with the 10.X multipliers and this CPU? Anyone else have this problem?

Oh well, so long as I can get it stable at 2.2GHz and beyond, I'm happy. Might have to get another one for my other PC.


One other minor issue - the CPU is recognized as Unknown CPU. Is Abit going to issue a BIOS update so that the NF7 series can recognize these mobile CPU's properly?

 

[Megatomic]

What BIOS version do you have installed? I have a mobile 2500+ on the same board and I have the full range of multipliers available to me. That was with D25 and D26 (stock not modded).

 

[Lonyo]

ABit haven't updated BIOS so far to recognise mobile CPU's, so I doubt it. An update may help the multi issue somehow though.
As for the other issue, no idea. Sometimes things don't like certain FSB's, but since you went over 200, and you're running 200 now (although with higher mult), it does seem odd it could have problems with some multi's (to me at least).

Megatomic, he says multi's are there, but it won't boot with them selected, even though speed and FSB are not the issue, so it must be the multi's having a problem.

 

[Megatomic]

I missed that comment but I am able to use the 10x mults. I used them when I was breaking in my new XP-M chip. I think he should try a new BIOS if he has an older one installed.

Regarding the Unknown CPU thing (not there when I replied...) it is normal. Hopefully Abit or some skilled modder will add XP-M CPUID namestrings to future BIOS files. I'm trying to get some of them to consider now.

 

[Jeff7]

Originally posted by: Megatomic
What BIOS version do you have installed? I have a mobile 2500+ on the same board and I have the full range of multipliers available to me. That was with D25 and D26 (stock not modded).

D26, just flashed it a few days ago.

 

[Megatomic]

Hmm, what did you have before you tried the D26? If you skipped over D25 give that a shot.

 

[Jeff7]

Actually, I had D25...but now I can't remember which came first - the BIOS flash or the bootup issues.
I might have flashed the BIOS pending the arrival of the new CPU, but it's more probable that I tried the BIOS update to try to solve the multiplier problem. I'm fairly sure the latter is the actual scenario that happened. But I could be wrong.

I might not bother going back either - it's been running Prime 95 since before I posted the first message - 2.2GHz, 1.6V, and no errors yet. That was with the full RAM test (Blend). I'll set it to maximum heat generation with the In-Place FFT test for awhile.

 

[Megatomic]

Jeff, I am able to do 12x200 at 1.7Vcore. Have you tried that combination yet? My system is a beautiful thing running at that speed.

If only the BIOS would recognize my chip. :heart:

 

[Jeff7]

Originally posted by: Megatomic
Jeff, I am able to do 12x200 at 1.7Vcore. Have you tried that combination yet? My system is a beautiful thing running at that speed.

If only the BIOS would recognize my chip. :heart:

Not yet; going to see that it does 11x200 at 1.6V first, then I'll go from there. It's already at 50C with an Alpha 8045 and AS5. But the AS hasn't set up yet, but that shouldn't do a whole lot for the temp. Case temps are good too - 35C. The Antec SLK3700 has great ventilation with just two 120mm fans..

 

[Megatomic]

I also use an Alpha PAL 8045 (with a Vantec Stealth 80mm) and AS Ceramique. My system is running inside an Antec Sonata with the stock 120mm fan and an Arctic Cooling VGA Silencer exhausing. At 12x200 at 1.7V my CPU idles at 49C with the system temp. at 31C. After 3 hours of Prime95 Blended testing the CPU is 56C with the system temp. at 33C.

I'm really satisfied with these results especially considering how quiet it all is. I've considered getting a faster fan but the system is completely stable as it is (and this is summer here in VA, too).

 

[Fern]

Jeff,

I'm using the original BIOS that came with my mobo and I was able to use the 10 multi (I was playing around trying to see if I could get to FSB 220 w/o the L12 mod etc.). Maybe I'm just lucky for a change.

From time-to-time, there are discussions regarding this and diff BIOS over at nForcersHq and Abit forums.

 

[loafbred]

I read quite a few posts on the Abit forum at AMDMB forums about inability to use the 10X multiplier. I think it might have more to do with the cpu used than with the BIOS, because I was able to get a friend's NF7-S to work at 10*200 with a 2400m Barton. I recall it having some difficulty saving that setting, but after one or two tries it worked. That same cpu is now running in my father's NF7 at the same 10*200, and I'm pretty sure it took a couple of tries before his BIOS would save that setting. Neither board had today's newest BIOS when using those multis, and I can't remember which BIOS version they were using. I do know that the NF7 is using a BIOS which some people were having 10X problems with. Both boards are rev 2.0 boards.

 

[Jeff7]

Bump, with a solution.