Powerd by meson

PIII Overclocking Problems, Solutions, & Results

by johnd - Feb. 10, 2000

with contributions by:
James Anderson, Clayton Brazil, Mark Hammond, Eduardo Chaves Rosas, Dylan Thomas, and Paul aka "Topshelf"

This article contains the results from overclocking 4 PIII 500Es and 2 PIII 550Es with 2 Celeron 300As thrown in the middle of it all. One of the PIII 500E CPUs discussed in this article was tested here at Hardware News Net. The others are the overclocking results of CPUs that some of our readers sent in to us in response to our " Overclocking 4 PIII 500Es" article. We discuss the problems that the users had in overclocking their systems and what they did to try and solve the problems. Thanks to everybody that has responded to the original article and shared their results.

CPU #3 Revisited

We originally had some problems getting CPU #3 to overclock past 675Mhz with any reliablity at all. While doing some testing with some video cards, we tried a Creative Labs Annihilator Pro with the CPU running on a Abit BF6 motherboard. Running Quake III timedemos with the system would occasionally give us some lockups and just didn't run as smooth as it should have with the Annihilator Pro. We had previsously overclocked the memory on the Annihilator Pro to 330Mhz on a Soyo motherboard with no problems. However, with this setup the system locked going into Quake III after setting the memory on the Annihilator Pro to 315Mhz. As suggested to us by James Anderson, we changed the SDRAM Leadoff to a setting of 4 to see if this would make a difference. Other readers suggested that there might be a voltage problem. I read where a reviewer of the Asus V6800 at SystemLogic had tweaked his AGP voltage up to 3.48v. This had cleared up some "picture anomalies" he had been experiencing with the card. After adjusting the BF6 to use a SDRAM Leadoff of 4 and setting the AGP voltage to 3.5v, we rebooted the system and tried our benchmarks again. The system still wouldn't overclock any higher with these new settings and still didn't run smooth while playing the Quake III demos, but we were able to get the Annihilator Pro up to 326Mhz on the memory.

Mark Hammond and others suggested trying older or newer versions of the BIOS. We had a recent version of the BIOS in BF6 board, but not the latest version that had just been released this month. So, we downloaded and flashed the BIOS with the latest available from Abit. After setting everything backup in the BIOS, a quick boot showed everything was OK. This time, instead of a Quake III timedemo that didn't run smooth, we got nothing but smooth. With the system now running smooth, we started adjusting the FSB. We moved the FSB to 144Mhz and attempted booting into Windows 98 at 720Mhz which this CPU motherboard combination had never done before. The system booted, ran Windows 98, and Quake III without stuttering one bit! So, the FSB was moved to 150Mhz for a 750Mhz boot, no problems. Again Windows 98 booted and Quake III ran. We then tweaked the FSB to 155Mhz and booted up. As soon as the Windows startup sound hit, the system locked.

750Mhz was over 30Mhz beyond what we had ever booted the system to before. However, after about 30 minutes the system locked up. The highest clock speed with a stable system on this BF6 is now 725Mhz, 50Mhz beyond where it had run stable before the BIOS upgrade. On this BF6, the BIOS upgrade was definitely a boost to our overclocking with this CPU. We did notice one curiosity, the AGP default voltage now reads 3.5v instead of 3.3v. It could be that Abit has figured out that the BF6's AGP bus needs a little more voltage to run reliably. I would suspect that this update may help users of the BE6-II also with their search for increased stability and faster overclocks, since the BIOS update for the BF6 is also the same update for the BE6-II.

Solution #1

James Anderson sent in the suggestion to try tweaking the SDRAM Leadoff under the advanced chipset features. Here's the situation where it had helped him out.

James was able to get his 500E to run with a 140Mhz FSB by switching his 2 DIMMs around in various combinations until he got a stable system running at 700Mhz. He suspected that his ATI Xpert 2000 AGP card was holding him back from going any higher. He replaced it with a Matrox G400, but instead of being able to overclock any higher, he started getting "little pink dots on certain toolbar icons in Windows Explorer and Outlook Express. Thinking the G400 was bad, he exchanged it for another, but the problem remained. He then found someone on the forums at www.matroxusers.com who had the same problem and eventually had posted a solution to the problem which was to change the SDRAM leadoff value. After changing it on his system, he no longer had any of the dots showing up, but he still couldn't overclock any higher. He decided to try his ATI card again. With the ATI card back in the system he was able to move the FSB to 150Mhz and run 3DMark 2000 without any problems. His scores from the G400 were "about 2 1/2 times better than the ATI card," so he put the G400 back in the system. Eventually, he replaced the G400 with a G400 Max. With the G400 Max installed in the system he was able to overclock the FSB up to 145Mhz, even going all the way to 152Mhz on the FSB as long as he didn't run any 3D apps.

Copyright 2000 - 2006 Hardware News Net. All Rights Reserved.  

page execution time : 0.016348 secs.

Powerd by meson content management system © 2006 John Dantzler