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:

LOSTCIRCUITS	SHORTCUTS: A new bus speed Heat and Hertz Test Configurations Cachemem, Sandra Everest and Speed Derating Sysnark2004 Internet ContentCreation Office Productivity WorldBench5 -1 WorldBench5 -2 3D Rendering 3ds max Cinebench2003 Futuremarks Gaming A-Z DOOM3 FarCry Conclusions and Paradoxes Give Us Some Feedback to Help Us Improve our Reviews
Intel Pentium4 3.46 Extreme Edition / 925XE chipset
(Review by MS November 16)

Conclusions

IN RETROSPECT, getting this review done – which included some 200 hours of running new benchmarks etc. – has left us with a feeling of déja vu bewilderment, some things at least on the surface don’t add up and it is in front of you but you just can't grasp it. At closer look, and after digging through the different benchmarks, some of our observations make more sense, however, not in the way we expected them.

The first issue is that the combination of the Extreme Edition with the 925XE appears a rather unlucky mismatch. The Northwood and, by extension, the Galatin core has historically suffered from a bottleneck at its front end. That is, in the case of the Canterwood chipset, a 75% bandwidth utilization was already a top score, however, without knowing better, we blamed most of this on some protocol overhead in the PSB interfacing. Increasing the bus frequency raises the bar but not even close to an increase in bandwidth that would scale in proportional fashion with the frequency delta. This leaves, once again, the front end of the CPU as the culprit.

Looking at the Everest memory READ benchmark, we see that that Prescott even on the slower interface of the old 925x host bus dramatically outperforms the Extreme Edition. Therefore, just from a logistics standpoint, it would appear to make more sense to pair the very same Prescott with the new chipset binning. Alas, below the 14 x setting, there are no provisions for lower multipliers on the Prescott and, at least with the last steppings, the resulting speed of 3.73 GHz would be prohibitive for a mass release, moreover, as we showed in this review, even at 3.6 GHz, the processor hits a ceiling somewhere.

None of this should be a final limitation, Intel just released the P4 570J specked at 3.8 GHz where finally some improvements in leakage current reduction were achieved, concomitant with the resulting reduction in idle thermal dissipation. If we look at all these seemingly isolated details, they suddenly fall into place and explain at least some of the results we have seen. Now it is up to Intel to adapt and match the correct pieces if they want to change the 925XE from an oddball to a successful chipset.

next page:     => More =>