John Loo never told us whether he has a Prescott or a Northwood.
That particular Northwood, at least with certain production runs, ran so cool that it would sometimes idle at a lower temp than the mobo, and you'd think something was wrong with your temperature sensors.
Of course, it could also depend on the mobo.
If it's a Northwood, then there is a REAL problem. If it's a Prescott, it still seems high unless he's upped his voltages to some insane level.
Alaa said he had a Prescott 630. What memory is he using? What kind of mobo?
I can see these high temperatures with the stock cooler, but I never used a stock cooler on my 3.2E. It has a TR SI-120. Even at room temperatures of 80F, the load under S&M or Prime95 never goes over something around 110F to 113F (43C). With an XP-120, my brother's 3.0C Northwood loads up to 118F at room temperature just above 70F.
. . . And that performance (on my Prescott) occurred with the VCORE set at just below 1.4V. I've since run it up as high as 1.4625V, and it only shows a 2C increase in load temperature.
I'd look at the heatsink retention mechanism and the thermal paste. If you're using a stock cooler, I'd also look into a better cooler like a heatpipe design. My 120x38mm fan spins at 2,200 rpm, and if you have a stock cooler with a little 70mm fan on it spinning at 1,500rpm . . . .well . . . . then I can see how these temperatures get up there on the Pressy -- 478-pin or LGA775 either way . . .
Mr. Loo also failed to mention if he over-clocked his system. But then, I've OC'd my own 478-Pressy, and I've explained my temperatures and cooling. Anyway, a 2.4 Pressy or Northwood should over-clock reasonably well with the right mobo and a good cooling system.
What other factors, the PSU for instance and its adjustment, would cause excessive thermal leakage from the processor cap?