xp2100 @ 2150mhz and 64kb of l2 cache what is going on

[bowman1964]

ok well i have a direct die cooler i hand made.runs -52f no load and -42f fully loaded xp2100@2150mhz at 1.95.but problem is now i lost my l2 cache.windows shows my processor as a duron with 64kb l1 and 64kb at l2.now if i let the evaperator warm up before booting sometimes it will show processor as a athlon at 2150mhz.but if i dont catch it just right boot will say unknown proccesor at 2150 mhz.if i can get lucky it will load into windows and run fine.has something to do with the extreeme cold because i can clock it down to 1500mhz or slower and still get unknown cpu.now i reflashed the bios and it reconized the cpu for a couple of reboots and then back to unknow cpu.i am wondering if maybe the extreeme cold is afecting my motherboard somehow. i have had the processor up to 2250 mhz but this isnt much good untill i get the l2 problem fixxed.
i hope someone has a anser.
system specs

xp2100 unlocked (agioa 0212 y)
xp333 volt modded
pc3000 will run 201 mhz and bench 512mg
radeon 8500 volt modded to 320core 340 mem

[N8]

As far as i know, this is pretty much a known problem with the Palomino XP's running under extremely cold temps. Think I saw somewhere that it was not such a problem with the tbreds. Sorry to hear this. But that is a sweet-ass cooling setup u have there :)

[sysfailur]

Yeah ditto nice cooling, and home made right? Extra coo!

[bowman1964]

YES HOMEMADE REACHES -40 IN LESS THAN 2 MINUTES.BUT THE CPU EVAPERATOR IS THE HI TECH PART OF THE SYSTEM.8 HOURS ON A CNC TO MACHINE THE DIRECT INJECTION MAZE BLOCK.

[Torinalth]

is there actually a change, or is ti just registering wrong. can you bench it and still rock all. or is there ACTUALLY missing cache? if it is just registering wrong with no ill effects, then I'd just not worry about it.

Torinalth

[bowman1964]

yes i actully loose the l2
3dmark2001se i lost almost a 1000 points

[DisposableHero]

sweet temps dude!

to fix, just buy a 2200+ ;)

[MrIcee]

I will offer one suggestion to try as I deal with this same problem myself on my system with an AGOIA 1700+. Although I am warmer than you at -30c...anything above 2.05v and I lose my L2. I will assume that the additional -15 to-20c is creating a problem that wherein even 1.95v, which I run with full L2, under your temps is still too much voltage. I would try reducing voltage by a couple tenths at a time and see if you can maintain your overclock but regain your L2. The only other alternative is which you already have explored, warming up the cpu...but in the long term you will not be able to maintain total cpu Mhz. Try reducing the volts in small increments and see what happens:)

Randi:D

[DisposableHero]

isnt there a registry setting or some file someone tryed a while back at icrontic that fixed this

[dmitriyaz]

i am sad to hear that...
i will be making a direct dye evaporation system myself,
i didn't know that VERY low temps are actually worse than low...

- are/were there any other CPU cores (beside Palomino) that behave strangely under temps that low?
- how common are things like that?
- is it the temperature, the high voltage, or the both together that cause this?
- how do people LN2 cool their PC's? wouldn't they have this problem too?
- even people with Prometeia's, their temps are around -40, why doesn't that happen to them (ot least not often)?
- does a similar thing happen when supercooling a video card, for example? or only CPU? (cause i am planning on direct-dye cooling a GPU, too...

if someone tried to answer these, i would greatly appreciate.
thanks

[OPPAINTER]

The XP2200 would be a good chip, it doesn't have this problem. if you look at the XP data base at VR Zone you'll see the Ln cooling guys kinda gave up on the Palimino because of this problem. They started back to AMD when the 2200 T-Bred came out.
http://www.vr-zone.com/guides/AMD/AthlonXP/
http://www.vr-zone.com/guides/AMD/TBred/

The way I kinda controlled the problem by lowering the voltage and make sure the Halt was dissabled, no idle. If the Halt is enabled it will mess it up.

OPP

[bowman1964]

thanks OPPAINTER i will try a little more tonight when i get home.i am glad you replied i have been following you and the guys for a long time.i am a old school extreeme overclocker from the old intell 300a days.but i am not much on the forums i have just got into them a little.and yes i am getting a new xp2200 but i need your best bet on a killer motherboard.i am going to pic up a new one thinking about the new epox.
and MrIcee thanks i have been watching you too .you guys are the best i have seen around the web.and ddung where is he latley.should call you all the 3 musketeers.
i think the icronticforums lost a great bunch of guys when you guys left

[bowman1964]

well MrIcee you were right on track.i lowered my voltage to 1.79 and reflashed my bios.now i got 2100 with full L2 but if i up the voltage much more i lose L2.but i have to clear the cmos to get the L2 back after the i raise the voltage too much.strange that i have to clear the cmos to get it back.but i am testing now so you helped me alot.

[KODE]

Hi bowman1964,

Well I went into the same problem you are having, but I found a very easy way to fix the L2 cache "bug". I noticed that in the first time I was booting my PC, It was showing "unknow CPU type 2048MHz" in the boot screen and I also noticed that every time it went into windows that way it was showing a duron with 64K L2 cache in WCPUID. Well I solved it just shutting down my PC at the boot screen and turning it on again (sometimes had to do it twice). After that the boot screen shows "Athlon XP 2048" (or something like this) and my L2 cache is full 256 again. Works like a charm!!

Resuming: When it boots "unknow CPU type xxxxMHz" =DURON 64K L2 cache. When it boots "Athlon XP xxxxMhz" = XP 256K L2 cache.

Good luck!!

Regards, KODE.