P4/2.4B overclocks nicely to 3.2GHz on air ~ would 2.4C do better?

[BCinSC]

Given the B starts at 533FSB and the C starts at 800FSB.

 

[TStep]

Given my limited experiences w/ somewhere around 4 or 5 2.4C's, 3.2-3.3 seems about average. I think 3.0ish was the least overclock and 3.5ish about the most overclock I've had truely stable.

Edit: the real limit in some cases is the high fsb demands on the mobo, and secondarily the memory. A P4B is much less demanding on the fsb and memory, but at a few % less performance.

 

[superkdogg]

Some steppings are better than others. Most C's will do just as well, many better. You'll need better ram and mobo to get them there, though because of the 133/200 FSB difference.

 

[BonzaiDuck]

T-Step's experience is more extensive than mine.

I've had a 2.53B and 3.06B CPU. I've also had a 2.4C and 3.0C CPU. None of the "B" (533 FSB) models over-clocked as well as the "C" (800FSB) models. The load CPU temperatures even at stock settings for the "B" CPUs always seemed higher than I wanted them to be.

Perhaps a needless word of caution: your motherboard should be capable of using a CPU with an 800 Mhz FSB.

Because I wanted to run a 1:1 divider, I never pushed my 2.4C beyond 3.0 Ghz. I was using DDR500 memory, so the external frequency setting was 250 Mhz and the FSB was 1,000 Mhz. I probably could have pushed the 2.4C farther, as T-Step observes. Right now, my limit for the 3.0C CPU is 247 Mhz external frequency and 988 FSB, or 3.0 @ 3.70. This is a self-imposed limit on exploring farther, because I am purposely choosing a particular trade-off between temperature and cooling solutions, speed and stability.

What I found amazing about the 2.4C (it was a 30-cap processor, also) was its idle and load temperatures. The idle temp was always equal to or below the motherboard temperature by maybe 2 degrees F, and the load temperature, as I recall, never broke 104F or 41C. Of course, at that time, those were my own home-grown load tests and not PRIME95, but I've since proven to myself that my own load tests came pretty close. And for the skeptics, I have since verified that my motherboard's thermal sensors are reasonably accurate.

I had achieved a "seemingly" stable setting for the 3.0C @ 3.808, but this setting would have generated tell-tale rounding errors under PRIME95. Maybe someone else has been luckier😉

ASUS P4P800 standard
OCZ EL "Gold" DDR500 1GB
Intel P4 3.0C @ 3.7
ASUS v9980 nVidia FX5950 Ultra

 

[BCinSC]

Did you 2.53 OC well? I've got a couple of those, too. The 2.4B is on a 845PE SOYO board. I'm actually surprised it's doing so well. I have a nice 865 MSI board. What does that get me?

 

[TStep]

Switching to an 865 board gets you the dual channel memory, so a few % of "free" performance if you already own two sticks of the same DDR. Hyperthreading would be the other difference between a B and a C.

 

[BCinSC]

845 board has RAID and I've got a nice pair of 300GB 16MB Cache drives.

 

[clarkey01]

I think Duive has a 2.4C @ 3.5 Ghz.

 

[BCinSC]

That's a seriously high FSB ~ 292MHz, which translates to 1167 in Intel's "Quad-Pumped" world.