Newbie OC'er needs help, Warning! Long Post*RESOLVED*

Fern

Elite Member
Sep 30, 2003
26,907
174
106
EDIT: As a preamble to a long post let me add , are the different problems enumerated below indicative of specific problems? Or are these just the consequences of an unstable overclock in general?

If they are indicative of a certain problem, that would surely give me a starting point on what to tackle.

In reading posts here and elsewhere over the past 4 or 5 months, I haven't noticed any posts detailing what happened when an OC wasn't stable. Ususally they just state it wasn't stable or failed prime95 etc. Consequently, I don't know if my unsuccessful attempt is the same as what others have experienced.

The only problem I understand (or think I do LOL) are those where memory errors are reported. OK, I'm thinking boost vdimm or relax timings.

Registry Checker probs and illegal operation warnings about MTASK and MSGSRV32 ? Huh, does my cpu need more voltage or is it memory or something else. I've no clue. Has no one else ever had those warnings etc when trying to OC?


ORIGINAL POST:

As the title suggests, I tried late last night to just go on up to 3200 (from 2500) by raising the FSB to 200 in one jump. See spec's in sig please. It will post fine, But chokes getting into Windows 98se. First I went to msconfig and (it was already at selective start) closed down everything under startup but systray. Then set mem timings at optimal (for stability)

Before any voltage mods, I got this error:

A fatal exception has occurred at 0028:C00322EB and 0028:C0032311

Don't know if that's a mem prob or system file corruption. Googled it, nothing came up. (EDIT: From searching the MS website, this appears to be a cpu-related error. Right?)

Anyway, bumped mem votlage a notch then got an illegal operation warning regarding mtask and msgsn32, hit enter and a blue screen came up with registry checker problems (restored my registry apparently). Hit enter then a warning about unreliable XMS memory @ address 0030F804.

And so it went until I got the volts to around (did'nt bring all my notes with me so at this exact point I'm going from memory) CPU 1.75 and mem to 2.8. Got into windows fine and could navigate around, ran scandisk and checked device mgr etc. So I ran Prime 95, failed immediatley with a rounding error. Took volts up again, same thing.

Think I ended up with mem volt @ 2.9 and cpu 1.8 maybe 1.85. Couldn't pass anything in Prime95, same immediate rounding error although the value (error) changed slightly.

The highest temps I got on the cpu were 20-21c (I'm live in the Smoky Mountains and the rigs in a very cold room). Didn't have anything installed to check mem temp, but the heatspreaders were cold to the touch.

Oops, forgot to mention I kept loosening mem timings till I was at 2.5-2-3-11

So, I'm thinking I will:
1) Flash up to the newer BIOS (never done that before, so I'll give it a shot if you guys think it will help)

2. Start by only going to say FSB 174, running Prime95 and if OK then jumping up by 4-6. I prolly need to start with either mem volt or cpu volts to isolate which is more effective/neccessary. Any suggestions on which to focus on first?

3. Run Prime95 at stock to see if windows is stable there. - already done this for about 20 hrs no prob (mem timing here is 2-2-2-5).


BTW, the whole time the network and raid features were on even though I'm not using it curently, any suggestions regarding that? I can see from FAQ's these s/b disabled in BIOS Correct ?


I'm on air cooling with a stock HSF. but since the computer is basically in a freezer, temps don't seem a prob (again, highest temp on cpu was 20-21c read off BIOS) Heck, I've even unplugged my case fans and just left the side off since its so cold here now.

Any thoughts are suggestions from you guru's would be appreciated.

Thanks

:confused: :confused:
 

Fern

Elite Member
Sep 30, 2003
26,907
174
106
Yeah, and the sad part is I didn't list ALL the probs. Now I'm just taking it in steps: the real helpful part is running memtest86 after FSB adjustments. Seems to help narrow down the probs (i.e., where to focus)
 

slaves123

Member
Oct 8, 2003
184
0
0
i think the problem is due to a bad overclock that's what tipically happens when the processor need more vcore
 

Fern

Elite Member
Sep 30, 2003
26,907
174
106
You Sir, are entirely correct. Anyway, I've decided to go back and take it slow. I started at FSB 174, and have been going up in increments of 4. After each increase in FSB I now run memtest86, then on to Prime95. That has helped narrow down the problems. As can be expected with my ram, the adjustment needed to gain stability has been a little vcore. Currently I'm at about FSB 190 with a vcore of 1.7 (as reported in BIOS), which I understand is prolly really only 1.65 as the mobo undervolts on vcore.

When I left the house this morning prime95 had been running about 15 hrs. Hoping all is well when I return home so I can bump it again :)