//flex table opened by JP

Click to See Complete Forum and Search --> : Danzego, the error magnet!!


Danzego
06-30-1999, 12:04 AM
Well, here I am again!! Been awhile since my last help request (thankfully), but this one is a big one. Lately, I've been the subject of all kinds of errors, be it blue screens or illegal operations. My system is as follows: ABIT BH6, PII 400 (SECC 2), 10 GB IBM Deskstar HD, Matrox Mystique 220 and Diamond Monster 1, Tekram 390F SCSI, AWE 64 SC, Diamond Supra Express 56K modem, CD and 4416s burner (Yamaha). Problem started when I put in the PII 400 and a 128MB stick of RAM. I already had everything else (I took out a Celeron 300a, non-oc'ed). I had a total of 192 MB RAM. Everything went ok until I yanked the 64MB stick I had and tossed in 2 more 128's (Toshiba ECC, disabled in BIOS). I kept getting the occasional blue screen and illegal operation in things like Kernel32.DLL, DIMM ID, <unknown>, and Shell32.DLL (all of these caused by Explorer). The blue screens I got were OE in VXD VFAT(01) and OD at 0246:013F47FB. I took out the original 128MB stick, thinking maybe it was bad or incompatible with my board and got more of the same problems (keep in mind with every RAM pull I start with a fresh W98 install for integrity's sake). Last night I put the original 192 MB back in the board after pulling the 256MB of Toshiba chips and first thing I got after install is a blue screen of VXD at VMM.32. I loaded the drivers for the Matrox 220 off of their driver disk to replace windows drivers. This morning I installed the latest drivers from Matrox for this card and couldn't even use my system cuz I kept getting illegal operations at startup (invalid page fault at 015f:00401f31). I couldn't even access safe mode to change the drivers because it got that error there too. Well, after crashing windows that last time after only 7 hours total use (and a couple of Windows automatic registry fixes), I decided to reload yet again, this time letting Windows keep it's default drivers for everything but my modem and SCSI (I also didn't load IE5 to see if that was causing the problem). So I come home from work today and the first thing I get when I boot is that **** invalid page fault at Kernel32.DLL. Also, when I try to clear it off the screen it keeps popping up again! After two restarts I got it this far so I can send this off. I've tried Nathan's help page to no avail (can't find my errors), so any ideas? Everything went fine until I put in the RAM and the PII400 (retail boxed), and I switched the RAM around everyway I can think of. It's not like the same things didn't work before, including Windows, and the error messages used to be few and far between. But 7 hours to crash W98? Plus one fresh boot 6 hours after reloading W98? What the deal, folks? TIA!! (sorry for the HUGE post)

R.I.P.
06-30-1999, 04:41 AM
Try posting over at Russ's board www.compucheap.com They know everything about hte bh6 there...

Ed_S
06-30-1999, 09:51 AM
Or maybe at the new one, "Hardware, Operating Sytems, and Troubleshooting"

http://www.hostclub.net/ubbcgi/Ultimate.cgi?action=intro

LJE2
06-30-1999, 10:04 AM
If you have Toshiba ECC RAM, why do you have ECC disabled in the BIOS? are you combining both ECC and non ECC? Also I have been told that on some boards ECC and non ECC RAM cannot be used at the same time. Try running just one type of RAM at a time.

[This message has been edited by LJE2 (edited 06-30-99).]

Danzego
07-01-1999, 10:18 AM
I disabled the ECC in BIOS in the first place because I was running them with the first 128MB stick which was non-ECC. I was told by a couple of people that as long as I had that disabled in BIOS I would be able to run the two together. I already have tried running the two seperate kinds alone. The time I crashed W98 in 7 hours flat was with the ECC RAM and this time I'm just using the stick of 128MB and 64MB together (both non-ECC). I'll post at those other forums ASAP. In the meantime, any other suggestions would be greatly appreciated. TIA!!

BBA
07-01-1999, 05:50 PM
If you've ruled out the ram, then look at the MB or CPU. The CPU is the likely culprit here.

What CPU markings do you have and have you ran CPUID? I bet you have a 2.0v 266 thats been remarked!

Try running it at 266Mhz with a 66Mhz FSB, see if the problem goes away! If it does, then bump up the CPU voltage to 2.1 or 2.2 and see if it works ok.

You can tell for sure if you check the current L2 latency setting of the CPU, I believe the setting for a P2 400 is 8 or 10, and a P2 266 latency is between 2 and 6. You can tell this by using the CPUL2 program.

BBA

[This message has been edited by BBA (edited 07-01-99).]

Danzego
07-01-1999, 08:04 PM
Hey BBA, is that a program bult into Windows (probably not) or something I get elsewhere? If I need to go download load it, toss me a bone or preferably a link. TIA!!

Nathan G.
07-01-1999, 08:22 PM
Here is your...good doggie bone.... http://developer.intel.com/VTune/cpuid/utility.htm
http://www.sysopt.com/forum/smile.gif

Danzego
07-02-1999, 01:34 AM
Thanks Nathan, but after using the CPUID, I couldn't find any other info on Intel's site to help me find out if this chip is remarked. It said Genuine Intel and 6,5,3 for the family, model, and stepping. Couldn't find if it was correct for a PII400. Still interested in that CPUL2 program though.... TIA!!

Danzego
07-02-1999, 01:45 AM
O.K. Now I'm freakin worried here. I used a program on my Maximum PC disk to check ID on a PII and I saw an L2 cache latency of 2. Are you sure about that spec BBA, and where can I find that type of info?

Nathan G.
07-02-1999, 08:39 AM
Hey man,
This is a link to TheTechZone software download page.
thetechzone.com/downloads.htm (http://thetechzone.com/downloads.htm)
On this page you will find Wcpuid 2.4,. It will show you current L2 setting (program is only 53K)
You can also find WCPUL2 1.4 on site. This app. will allow you to adjust L2 latency above 3.
FYI=my L2 latency is set to 3 on my P2-300 (SL2W8) @ 504.
I do not believe you can relate L2 latency to CPU design speed. I have 2 of the same stepp,spec SL2W8 CPUs. One has 4.4ns and the other has 5.0. Both are p2-300, so stepping and L2 latency must differ between CPUs with same S-spec.
AS for ensuring CPU is real , go to this link at intel and select the steping of your CPU. (in my case SL2W8). and compare data with you results on CPUID this should let you know if its the "REAL DEAL". uhh but a good re-mark has proven to fake this out!!! sorry
BTW,I have never seen or heard of a re-mark in the SECC2 package, just the SECC type .
developer.intel.com/design/pentiumII/qit/ (http://developer.intel.com/design/pentiumII/qit/)
Hope this helps
really glad its FRiday,Have a great one!
Nathan G.


[This message has been edited by Nathan G. (edited 07-02-99).]

BBA
07-02-1999, 01:42 PM
Actually, I'm dead certain about the L2 latency being set different in an otherwise identical P2 or different speed, but only for sure with older CPU's (SECC)

There is even an article about it (or was) on Tom's hardware page or on AnandTech.

Basically, Intel knew they had a L2 problem, so they set the Latency differently on teh faster cpu's because it just wouldn't work at 400/450Mhz with the lower setting.

This is part of the reason a P2 266 deschutes will outperform a P2 400 if both are ran at 400, the catch is you need to be lucky and get one with cache chips that will perform.

Yes, Intel does post the L2 latency settings in their design papers (white papers) and this is available on their website, you just have to search for it and it clearly states the different L2 latency settings for each CPU speed, since they can maximize performance on all speed range CPU's.


If your default L2 Latency is at 2 and it is not a remark, then Intel has been playing with their cache chips in the SECC2 range.

Check out this table for stepping info: support.intel.com/support/processors/sspec.htm (http://support.intel.com/support/processors/sspec.htm)

According to this page, it is either a P2 350 or a P2 400, It may not be a remark.

BBA



[This message has been edited by BBA (edited 07-02-99).]

Nathan G.
07-02-1999, 03:52 PM
BBA- interesting. Is this a funtion of the TAGRam chip??could you please provide me a link to the Intel white papers that show L2 latency setting in relation to cpu speed? I have looked and cant find it. please provide a link to the PDF file and a page #. also a link to AnandTech article if you could
Thanks http://www.sysopt.com/forum/smile.gif
Nathan G.

[This message has been edited by Nathan G. (edited 07-02-99).]

Nathan G.
07-03-1999, 07:21 AM
BBA- I am not trying to be difficult,just want to learn. I can not find the Intel application notes that "clearly state" L2 latency for each CPU speed. please point me in the correct direction.(insert joke here)
Have a Great 4th!!
Thanks
Nathan G.

[This message has been edited by Nathan G. (edited 07-03-99).]

BBA
07-04-1999, 12:05 AM
Look here for some tidbits:

http://216.92.13.166/releases/98q4/981215/overclock-pii300-04.html