All Intel(r) CPUs are currently clock multiplier locked, yet FSB speed remains unlocked for the time being. This means that as long as the user chooses the P3-450's 4.5x clock multiplier setting, they can select and experiment with any of the available bus speeds that their mainboard offers.

Here's a look at all of the "possible" overclocking combinations that the Celeron(tm) 400 could see on an Abit BX6 2.0 mainboard.

Stock:		4.5 x 100MHz	=	450MHz
		4.5 x 103MHz	=	464MHz
		4.5 x 112MHz	=	504MHz
		4.5 x 117MHz	=	527MHz
		4.5 x 124MHz	=	560MHz
		4.5 x 129MHz 	=	581MHz
		4.5 x 133MHz	=	600MHz

The P3-450's lightweight 4.5x clock multiplier provides a wide-range of friendly speeds in the CPU's sweet spot, and we experimented with all of them.

We utilized a standard OEM AAVID Pentium(r) III fan/heatsink for all of our overclocking tests, in lieu of a new shipment of advanced P3 coolers that's being sent to us by our principal supplier next week. Once the new prototype coolers arrive we'll put a variety of SECC2 CPUs through the overclocking meat grinder and produce an article based on our experiences and results.

Our P3's AAVID fan/heatsink was very basic, we consider it a poor model for most overclocking applications as its heatsink doesn't come into contact with the P3's exposed L2 cache modules, and it provides just a single center mounted fan to disperse excess heat.

What shocked us was the ability of this early P3-450 to overclock stabily at very high frequencies without radical cooling measures.

Here's how it panned out:

	4.5 x 103MHz	=	464MHz	=	100% Stable
	4.5 x 112MHz	=	504MHz	=	100% Stable
	4.5 x 117MHz	=	527MHz	=	100% Stable
	4.5 x 124MHz	=	560MHz	=	100% Stable
	4.5 x 129MHz 	=	581MHz	=	Fail
	4.5 x 133MHz	=	600MHz	=	Fail

The P3-450 needs very little prodding or voltage tweaks to surpass a speed of 550MHz, our particular example endured four straight hours of Brett "3 Fingers" Jacob's Quake2 Crusher Timedemo, as well as four hours of Prime95's "Torture Test".

We don't have enough examples of the P3-450 yet to predict if this type of overclocking supremacy will be the norm for the line, but we're impressed by our early results.

581MHz failed to boot into Win98, after POSTing without a problem. We suspect that with the right type of cooler 581MHz might be possible for some of these CPUs, we'll know more when we receive more P3 CPUs for evaluation next week, including three 500s.