512K L2 Barton cache not showing up in memtest86

[Wolfchild]

Okay so I got a used unlocked Barton 2500+ and A78NX motherboard
and promptly cranked it up to 3200+ speeds (11x200).

I'm running memtest86 on it and didn't notice this earlier but
memtest86 reports the 128K L1 cache correctly...but says "unknown"
for the L2 cache.

Is it possible that memtest V 3.0 just doesn't detect cache
properly in a Barton....or is my L2 cache in the processor hosed?

No errors at all and it seems to be running stable and cool.
This is all just running on a test bench for now...when I install an
OS I'll see what Sandra says. Only thing that worries me is the
previous owner had the VCore up to 1.73 and I ran memtest
at the 3200+ speeds on it for a couple hours like that. Turned it
back down to default when I found out about that, and it's still
running memtest fine, but like I said, earlier I had not noticed
whether it was reporting the L2 cache or not.

BTW yes the L2 cache is enabled in BIOS.

 

[Sust]

hey wolfchild,

Ive currently got your same symptom with my barton 2500.
Did you ever get any definitive answer as to why memtest 3.0 is not showing the L2?
My only problem now is that I cant get my leadtek k7ncr18d-pro to finish memtest when I bump FSB up to 200MHz with 1.75V at the Vcore.
I even have the memory timings set to "optimal" which is leadtek's way of saying sloooow.
It ALWAYS dies at the beginning of test#5 and requires me to press the "power off" button in order to get something to happen.
Once you turn it on again though, it's as if nothing had gone wrong and even boots into winXP.
It is currently in its 14th hour of SETI@home running at 200MHz FSB.

 

[Wolfchild]

I did some searching outside of this forum after posting this thread and it turns out that others have the same problem, so there shouldn't be anything wrong with my L2 cache. It is showing up in BIOS too BTW. Just a limitation of memtest, maybe a new version will correct that.

Test #5 causes some errors with mine at default voltage and 11 x 200, but at 1.725 it ran memtest all night long. Before I install the board I'm going to see if I can go even lower on the voltage. Never had it lock up like that though.

 

[Sust]

Yeh, it locks up at the beginning of test 5 on 200MHz FSB, but when I bump it down to 166MHz everything works fine.
I can even crank up the memory timings to the max and memtest refuses to die.
*sigh* Maybe this is just a limitation of the board's innate capabilities.
It's a leadtek K7ncr18d-pro and it doesnt have "official" support for 400Mhz, but I hoped I wouldnt have too much problems hitting that speed.

Well, I shouldnt say that I've necessarily failed.
If seti@home is still crunching away when I get home, then I at least know the CPU can handle 400MHz FSB speeds.
Well, onto the albatron board from newegg.
Hopefully, Anand is right and it turns out to be a better overclocker than leadtek K7ncr19d-pro.

BTW, have you ever touched your northbridge heatsink after a few loops of memtest?
I originally thought that my northbridge chipset was getting too hot so I plopped a makeshift fan on there and it still died on test5.
Just wondering if potential northbridge cooling inadequacies could be contributing to my overclocking hiccup too.

 

[Wolfchild]

No my chipset heatsink doesn't appear to be getting very warm at all...
has pretty big passive aluminum fins though. And this cpu hsf kind of blows
some excess air at it.