Via Apollo Pro133A @ 150 mhz without VIO mod?

[Michal[mt]]

Hi ALL! ;)
Today I have made an experiment...I overclocked a bit more my Celeron2 800 from 1120 mhz up to 1200 (150 mhz FSB).
I have to say that my celly is pretty going @ 1120 mhz with a 1.95 core voltage, all seem to go well: graphics, games, internet, 2d sheets and more...
But...there's a but :rolleyes: when I try to make my cpu working @ 1200 mhz only 2d graphics wants to work...3d graphics in games like Unreal, UT, Severange and others is corrupted...it seems that the 3d transform process is affected by some bugs...maybe caused by the too high frequency? :eh?:
To made my cpu stabil @ 1200 I have increased the vcore up to 2.1 volts and in this moment I am writing to you all @ 1200 mhz with a bus ratio FSB/RAM of 1:1 ...so my infineon dimms ram are going @ 150 mhz.
I tried if the problem was caused by the too high ram settings and I checked it wasn't. I downgraded the ram frequency to 112 mhz and changed the cas timings from the more aggressive 2-2-2-5 to the standard 3-3-3-6 and nothing changed: the 3d grapchics problem did not disappear.
So I don t know wich the source of the problem would be...maybe the Apollo pro133A chipset can t be stabil @ 150 mhz(I got an Asus P3V4X mobo)...maybe it needs a VIO increase?
Maybe my cpu can t reach the awesome 1.2 Ghz speed? ...is it possible when some people in the database say they have reach this frequency with no problems?
...I forget to tell you that my cpu is a cD0 stepping (SL54P) and that cpu temperature seems not to be a problem, infact I am using a thermalright SK6 with a black label delta fan.
Thanks for any suggestion guys.

[KILLorBE]

You could try to cool the clock generator, it seems that it helped me a bit, but I also think my chip just wont go higher (FC-PGA cA2 PIII600@888 100% stable, @900 I get some errors).

After cooling the clock gen. it runs a bit more stable @900 w/o upping the CV (2.23V before, 2.13V now).

Here (http://forums.overclockers.ws/vb/showthread.php?threadid=53969) is the thread about the clock gen. cooling.

Upping VIO might help, but my P3V4X is already running @ 3.5 VIO (should be 3.3) so I don't feel comfortable running it @ 3.8 VIO.

[funnyperson1]

try setting the AGP bus to 1/2 using the dip switches, and can you adjust the VIO using this board???

[Michal[mt]]

Hi KillorBE :)
Very very THANKS for your reply but I have just made this work :confused:
Can you tell me which kind of errors did you get?

[Michal[mt]]

Hi funnyperosn1 ;)
Yes...my AGP bus is just configured to be 1/2 of the FSB (so in my case it works @ 75*X)
...I got a jumper to switch from VIO 3.5(standard for the Asus p3v4x mobo) to 3.68 but I am not sure to make this...KillorBE also sayd it will not be a nice choice.

[funnyperson1]

is this listed in the manual cuz i havent seen it...hmmm i learn something new every day...just too lazy to take my cover off the case...

[KILLorBE]

Originally posted by funnyperson1
is this listed in the manual cuz i havent seen it...hmmm i learn something new every day...just too lazy to take my cover off the case...

No its not in the manual (at least not that I know) the jumper is close to the CPU_FAN header. (There's some fine printing on the edge of the board).

Michal[mt] I got some explorer errors (usually means your system isn't 100% stable).

[Michal[mt]]

Thanks for your help guys ;)
...the vio jumper is not documented but it is situated near the fan plug...about your errors KillorBE...I have to say that I got them when the cpu temperature become too high, too(I got the Mstack error on startup when I reset the machine after playing Quake3 for example.). My final words are: the problem is the cpu temperature ;)
Funnyperson1, KillorBE, do you think the same thing?
Maybe putting a peltier I will solve this issue?...

[funnyperson1]

well i dont think so unless your heatsink is not seated properly or you arent using thermal paste because an SK6 should manhandle a Celeron, the SK6 can handle Athlons up to any availabe speed, so a Celeron should not be a problem.......what are your temps anyway, they should be below 45 at full load....

[Godfodda]

To answer one of your original questions, I'm pretty sure the Apollo133A chipset can handle the FSB easily. Many of us with Abit boards using this chipset are hitting that or more (see 866 machine in sig).

If it's only 3d that's not working, I'm thinking that it's vid card related. And I can't find where you posted your card model. I've read that some don't like to be too far out of spec. My Voodoo5 is happy at 85 but she's pretty warm at that speed.

Also, what are your CPU temps? I'm guessing the SK6 is more than adequate for your CPU, but if the case temp is high then the CPU temp will be high too, thus causing problems.

[Michal[mt]]

Still very thanks guys...
Mhhh...infact my temps are so hot...idle temp is 31°c and full load is 62°c. Am using motherboard monitor as temp monitoring software. This software directly indicates the temperature of the internal cpu sensor.
About the thermal paste...am using Artic silver thermal paste...what is wrong?
My graphic card is a KyroII(hercules graphics 3d prophet 4500)...but how I sayd before Godfodda, my system indicates some errors at the startup if I reset the machine after an intense game playing.

[funnyperson1]

umm thats way too high, but it shouldnt be with an SK6 , when you applied the Artic Silver did you scrape the thermal pad off first??? take the heatsink off, clean off with a lens cloth and acetone the bottom of the heatsink, apply a thin layer of asIII with a credit card and then reapply the heatsink making sure it is making proper contact (hint the notched side sould be on the part of the socket that sticks up over the chip)....sounds like youre having contact problems...

[Michal[mt]]

Hmmm....have made all you have sayd(the first time)...All the times I pick up the heatsink I clean all the components very well.
The strange thing is that 15 mins ago I have disconnected the SK6 from the celeron surface to make a simple control...well I have cleaned all parts and then used the silicon grase instead the wanderfull Artic silver...what happened? My idle and full load temps are 1-2°C lower.
How is it possible?
Another thing I want to ask you Funnyperson1 is which kind of cpu temp software do you use? Personally how I sayd before I use MBM (motherboard monitor).
Thanks again.

[outhouse]

I hope your temp sensors are wrong because your way to hot. Your temps from idle to full load should not be that far apart i think mine is 4C now its 20C and 23C-24C under full load. You may know this but when you overclock temps are even more critical then normal, if my CPU hets to 33C it will lock up rite quick. I'm curious about youir case cooling? Hers a few things ive found about high FSB settings and the Apollo 133a. #1 use a big HSF on your northbridge. #2 make sure sideband is disabled in your registry. #3 if any MB component or chip is hot put a HS on it. #4 and the most important keep everything cool but at 31C idle temps you are going to have your hands full and i wish you luck. Oh remember to run at the lowest VIO setting you can once you get your FSB maxed at 3.7 you will be doing a bit of gambling.

[Michal[mt]]

Thanks Outhouse!
You see...just now have learned more about my wanderfull cpu and have to say that inside it it doesn t have a thermsistor...so my temps are obiviusly given from an external diode.
Secondly I think you re right when you say to turning off sideband addressing...I remember...ehm not well ;) I read some time ago that this feature caused some corrupted graphics error when overclocking cpu's.
So...I am going to disable this feature from my bios registry and then post results.
About my case cooling? Nothing ;) ...it is only open.
Stay tuned...
:p

[outhouse]

I think MBM reads from within the CPU itself [internal diode] you may want to try sisandra and see if your results are different allthough without a way of knowing temps i guess your still in the dark.

here's what ive done to keep case temps down [20 bucks.]

First cut all the fan grills out of your case to improve airflow, use 1 intake and 1 exhuast and makle sure your PSU fan is getting 12 volts and running at full RPM. If your HSF for your CPU blows air down on HS then make sure your rear case fan blows cool air in and the front fan out, this makes sure your CPU is fed with cool air not the warm air as it is being exhuasted. I have made a duct that goes from my sidecase cover to my HSF and this dropped temps allot way more then an open case and this mod is free. Another free thing that helps a little is to round your flat ribbon cables, you can cut every 10 wires [and stack] for your floppy and cd but its best if you just fold your hard drive cable. Zip tie all other wires for improved air circulation.

well just ideas have fun

[funnyperson1]

all PIIIs and Celerons and P4s have internal diodes, i thought my temps of 65C+ were wrong but then i switched heatsinks and my diode was frighteningly correct, also the P3V4X does not measure using external diode...

[Michal[mt]]

I won t never tell you all thanks!
...but I have to tell you some bad news:
After have disabled "side band addressing" nothing changed, 3d corruption error still gives me his wanderfull glitches. :eh?: (only Quake3, sacrifice, and Max Payne are not affected by this problem...what to think?)
Have reapplied the Artic silver paste after well cleaned the two cpu and heatsink surfaces and MBM still gives me the same temps.
(idle: 29-31°c -- full load: 62°c)
Have tried if sissoft sandra tells something different about but nothing...sandra says 36°c after launched the "mobo information" function.
Yesterday I sayd mine cpu didn t have a thermsistor inside it...I was confused by a "otion string" from the Cpu-z program...so temps are right? I'm asking to myself how it would be possible to reach that temps with a nice cooling system like this and with the case open...
You all guys are so pretty...but I am becoming suspicious and thinking that this mystique ;) cpu never let me to set it @ 1.2 Ghz.
Ah...I wanted to let you see some pics I have done from my system and glitches errors(some of them are integrated here as img code, others are links to images probs)
I have noticed two glitches in the windows environment...here they are:
http://members.xoom.it/Michal/problem0.jpg
http://members.xoom.it/Michal/problem1.jpg
The second one is noticeable by a white plot near the basket.
These other links are to the UT graphic corruption image and to the sissoft sandra cpu benchmark.
http://members.xoom.it/Michal/problem2.jpg
http://members.xoom.it/Michal/sandrapic.jpg
I have to say another one time that games like Quake3, Max Payne, Sacrifice don t give this error.
Have tested the cpu in the short time(only 2 hours) with the Prime95 program and all went fine...no crashes.
Guys...don t know what to do...I think a better air circulation surelly will help to downgrade the temps of 2°c at least but I think this cpu never will go higher than 1.120 Ghz...

[outhouse]

You can try to burn it in at a stable speed for a couple of weeks [with the same CV your at now] running something like Seti 24\7 then there will be a good chance that you may be able to squeez a few more MHZ out of her as time passes.

[Michal[mt]]

It is all strange...have made Prime95 torture test and all went fine but still have this graphic glitches...Could be my beta graphic card bios? Bah! ;)
However very thanks outhouse, thanks you all guys! :beer:
I promise...if I ever will make this cpu 1.2 Ghz working with no probs I ll posts my results.
Bye all :burn:

[6502kid]

After reading your thread I got 3 suggestions.

Since you can run quake, but not UT, perhaps DirectX
has gotten corrupted. Try reinstalling it.

If you are running at AGP 4x, try 2x. It might fix the
problem, and you wont lose much performance at all.

Last, that CPU temp is way to high. Not sure what to
tell you there that you haven't tried.
Perhaps a new HS and fan, and some case fans wouldnt
hurt either.

Good luck.

[outhouse]

not only that but he's got a point i had some bad beta drivers that gave me the same problems he has told us, try the official ones and let us know how it works.

[Michal[mt]]

Here we are again ;)
Ok...Have tested for 12 hours my cpu @ 1.2 Ghz (2.0 vcore voltage) with Cpuburn in high priority mode and the result was: no errors.
But I have to say that you are right...my temps are too high and maybe this cause some artifacts in some overclock sensitive games. Just today have read a nice article that was talking about some problems that UT gives with some overclocked system...this article sayd that UT is the most overclock sensitive game and it is very easy to see some graphics artefacts with a very overclocked cpu.
The only way to solve this issue is to implement a better cooling system...so I am intentioned to buy a 85 Watts peltier and to watch what happens :p ...hope this will solve my problem.
...have tried if the problem was caused by the beta graphic bios or by graphic drivers and I checked they were not.
The only solution is to make temps lower.
I am very glad for your help guys...you helped me a lot...Thanks again! :beer: