Navigate: Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home	.	.
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ Search Prices:

After a bit of a slow-down, that is, no new releases of new speed grades in over 7 months, Intel finally catches up again with lost terrain towards abiding by Moore's law. The latest speed grade is 3.2 GHz, running at a 16 x clock multiplier and 800 MHz PSB data rate. Geared towards the Canterwood chipset, and performance acceleration technology, the 3.2 GHz CPU takes full advantage of the extra cycles to unleash brute force to the desktop.

Is this the final version of the P4 as we know it? Is there still a chance to push the system performance a few notches higher using the existing infrastructure or will any future speed grade only provide "empty magnification"? Along the same lines, what will be the most likely next steps in the Intel camp? We have some ideas about that, too.

The wafers have turned once again and a new speed grade of P4 has been released, this time pushing no less than 3.2 GHz. Otherwise, there is nothing new, no new technology, no new packaging no new functions. The most striking improvement is the new heatsink, finally this one's clamps are actually working the way they are supposed to work and it actually even looks pretty nifty.

A new cooler for the new speed grade

Theoretically we could end the review right here. It is not that simple, though. Increasing the PSB to 800 MHz data rate has reduced the multiplier to a virtual 3.75 x (15x / 4) for the 3.0 GHz version and, hence opened up some more head room, further assisted by the use of "Dual Channel" DDR3200. Introducing performance acceleration technology has further reduced access latencies and the combination of both bandwidth and low latencies has pushed the performance way beyond anything that could have been extrapolated from the pedestrian Willamette at the introduction of the P4.

In any system already pushing the limits, performance will scale rather poorly with CPU frequency increases. Bottlenecks within the system are becoming the overall limiting factors, irrespective of whether it is the graphics card or the memory or else the I/O interface, a.k.a. HDD. With respect to the latter, Intel certainly addressed the imminent threat of running into a brickwall with the release of the ICH5-R. Admittedly though, the solution at hand appears still somewhat immature, both with respect to the lack of configurability (no RAID1) and the more blatant omission of driver support for anything but RAID 0. However, as we mentioned just above, it is not a question of stability or support of the dying Parallel ATA platform but rather to maintain a certain amount of platform scalability with upcoming processors such as the one we have in the crosshair today.

next page:    => Test Platform =>