Slot 1 P3 700 cB0 problems at 125+ Mhz

[Muggy]

I've been having a lot of fun tweeking the last ounce of performance out of my P3 700 system. I'm currently running at 869 (124 fsb) but can't seem to get it any faster. I can o/c it to 933 with no problems as far as windows is concerned. I do have to up the voltage a bit (1.75) but it runs pretty decent.

My problem happens on all fsbs of over 124. Windows loads fine but if I attempt a benchmark (I happen to be using Unreal Tournament for this) or play an actual game, the system locks up. No BSOD, just hard lockup. I have to do a hard reset in order to recover. Does this sound like a memory issue? I've checked my memory with the latest version of MBM and it reports an operating frequency for my memory as high as 125 Mhz (and that's with it clocked at the normal 100 Mhz fsb). This is Micron memory of the 7.5 ns variety and should do 133 cas3 with no problems. When overclocked at the 124 fsb setting, I've actually got my memory set at cas2 and it runs rock solid.

My other thought is an out of spec AGP prob. Any thoughts or suggestions are greatly appreciated.

BTW my current temps and such are:

ambient 28
Case 30
CPU Idle/Full 36/42

The CPU temps are a bit high for a VOS32 from what I've seen. Lapping?

[batboy]

What motherboard are you using? Does it have a 1/4 clock divider at 133 MHz FSB? If so, everything should be in spec at 133 FSB. My guess is you have an old BX mobo and the network card is crappy out at high FSB. Also, those temps, while not too bad, are a little bit high. If you could get the temps down a little lower, I'd try raising the voltage up one more notch.

[Slake]

Same problem with my FCPGA 700E. Windows apps run fine at 133/133/33. 3D locks the system after a couple of minutes. No luck running the memory at 100MHz either.
Step Thermodynamics PC 133 ram- 128MB (running Cas 3)
Asus CUSL2-C Bios 1007 final
Vodoo 3 2000 PCI
2 NICS
Idle 32C Load 42C
Strange, I am running 3d games stable at 131/131/43
No shared interrupts. I disabled com1 and com2 freeing their IRQ's to avoid conflicts.

[Muggy]

Yeah, I totally forgot to mention motherboard and a few extra things. Motherbaord is an ABIT BX6 r2. My clock divider is set at 1/4 for the 124 fsb (31mhz PCI). The mobo offers fsbs of 127 and 133, plus the numerous other settings above that (which will remain untried till I get the kinks worked out).

Like I said, with the 124 setting everything runs great! Keep in mind this is PC-100 memory running at 124 CAS2 settings (2-2-2)! I can't see any reason why it wouldn't go at 127 CAS3 so I've pretty much ruled out bad memory. I am running 3 DIMMS (128 each) and I may want to experiment with 1 chip at a time. I'll update this thread if it changes anything.

Also, in response to your suggestion, Batboy, I have run the voltage up to 1.8 on the next fsb setting and I got the same results. Come to think of it, same temps too. I don't think I clarified that the crashes happen only on 3D apps. So let me throw that into the equation as well.

On another note, I did have some problems with my VOS32 and how it mounted to the proc. It didn't want to sit flat on the proc die, so I took the advice of another review I read somewhere and stuck a penny under the clip for extra tension. That seemed to help but I'm still a bit leary about how the HSF sits on my CPU.

[KILLorBE]

I would suggest you try to up the CV to 1.85 or maybe 1.9

[Muggy]

That's a thought. Although it doesn't make much sense that it'd run great at 869/124 1.65V but crash at 903/129. Come to think of it, I do believe I did try the 1.85 setting and the same thing happened. I'm about ready to think it's an out of spec AGP that's causing the problem. I can't change my divider to anything other than 1/1 or 2/3. That puts my AGP clock at 86 (129fsb) instead of 82.6 (124fsb). That's still such a small change that it just doesn't make sense. Any other suggestions? I'm still at work so I can't try this stuff yet. I need about 4 or 5 different options to try when I get home! BTW: I see I put 127 in my last post, that should be 129. Thanks a gain for the replies!

[Slake]

I'm not so sure you can rule out the memory as the culprit. I've had a few very experienced members suggest my PC 133 just ain't up to the task. Maybe you could borrow some higher rated memory and try it at the higher FSB. 124MHz may be the ceiling on your PC 100.

[Muggy]

That's probably the thing I'll try first. I've got a good friend with some PC133 CAS 3 memory on an 800 MHz T-Bird rig. I'm sure he's running 133 fsb and it's incredibly stable (not o/c'd). I'll get back with you all later on this one. You know, I wish there was some sort of published guide about what problems result from what components. Unfortunately, the whole system is so synched that any component can cause the same problem. Oh, well.

[DaveB]

When I built my PIII 700 cB0 system a year ago, every time I hit a hard lock in 3DMark2000, extra voltage did the trick. I needed 1.85V to get stable at 933 MHz, 1.9V to get stable at 970 MHz, and finally 1.95V to get stable at 994 MHz. It's been running rock stable at 970 MHz+ for over a year, 994 MHz since October.

You should definitely run that PC100 ad CAS3 until you can get some CAS2 PC133. Avoid that 124/41 setting because you might corrupt you hard drive. I don't like i815 boards because 1/4 divider doesn't kick in until 133 MHz exactly because of those the stupid ratio settings. On my Soyo BX board, it kicks in at the 124/31 setting. You really need to get to 133 MHz on your Asus board.

Your Radeon should have no problems at 90 MHz, so I don't think that's your problem.

[samuknow]

I am getting the same prob. I have a Soyo 6VBA 133 and a PIII 700 cB0.

Once I get close to 129 FSB she locks up tight in games. If I set a jumper on the mobo for 133 FSB my AGP divider will go to 1/2 haven't tried it yet.

I am sure that my AGP is to outta spec and hence the lockup. I might guess the same in your case.

Plus might look into some 133 mem. A good stick of 133 cas 2 is about 30$ right now.

I will check back to see what you find out.

[Muggy]

DaveB (Jul 12, 2001 03:12 p.m.):
When I built my PIII 700 cB0 system a year ago, every time I hit a hard lock in 3DMark2000, extra voltage did the trick. I needed 1.85V to get stable at 933 MHz, 1.9V to get stable at 970 MHz, and finally 1.95V to get stable at 994 MHz. It's been running rock stable at 970 MHz+ for over a year, 994 MHz since October.

You should definitely run that PC100 ad CAS3 until you can get some CAS2 PC133. Avoid that 124/41 setting because you might corrupt you hard drive. I don't like i815 boards because 1/4 divider doesn't kick in until 133 MHz exactly because of those the stupid ratio settings. On my Soyo BX board, it kicks in at the 124/31 setting. You really need to get to 133 MHz on your Asus board.

Your Radeon should have no problems at 90 MHz, so I don't think that's your problem.

Thanks. My board is actually an ABIT board. I am running the 1/4 divider at present. I don't see much use in running the 1/3 even though it is offered.

[Muggy]

Slake (Jul 12, 2001 10:29 a.m.):
I'm not so sure you can rule out the memory as the culprit. I've had a few very experienced members suggest my PC 133 just ain't up to the task. Maybe you could borrow some higher rated memory and try it at the higher FSB. 124MHz may be the ceiling on your PC 100.

Success! I was able to try some PC133 CAS2 Memory (Crucial) and everything went well at 129Mhz 1.65 CV. I tried the next setting of 133 and Windows wouldn't boot. So I did the logical thing and upped the CV to 1.75. After that everything ran fan. There was only one small problem. The memory I got was only 64MB!

What's even stranger is that when I put the PC100 stuff back in, one stick at a time, it worked too! I stuck with 256 MB since the 384 I had before wasn't working. I think it's very likely that my three Micron PC100 128MB sticks didn't like doing the o/c thing very well. Two on the other hand, seem perfectly fine.

I'm happily running at 933 now, and am very grateful for all the suggestions and pointers! My signature has been updated to reflect the changes.

Oh, and my temps are right around 40C full load. Is it worth trying to get those any lower? My ambient room temperature is always pretty high. Internal Case temps run about 27C so I think the full load temp looks alright. I'm open to suggestions though if anyone feels these are a little high. Thanks again.

[Muggy]

I forgot to change the fsb in my sig. Hope nobody thought I'd found a way to hit 933 at 124 Mhz with a 7X multiplier!

[samuknow]

Very cool.

I am going to try some diff memory. See if that helps..

[DaveB]

Yeah, I guess I got the Asus CUSL2-C from Slake since you didn't list your mobo. A minor brain fart on my part.

Anyway, I've been using an Alpha PEP66 on my FC-PGA PIII 700/Iwill Slotket II combo and it keeps the temp at 34C max under load when running at 994 MHz. So, yeah, you could do better in the cooling dept. But your BX6 R2 isn't known for going much beyond 133 MHz since it's an older BX board. The newer Abit and Soyo BX mobos can go to 150 MHz+ if your peripherals are up to it. But you'll never know 'till you try. Both better cooling and a higher core voltage will be needed to get further. Also some older Abit boards had problems running at exactly 133 MHz. You might try the next highest FSB just for the hell of it.