|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
|
Prices: 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 |
LOSTCIRCUITS
|
|
| DFI 855GME-MGF and Intel Pentium M 735 Along came a Dothan | |
|
(Review by MS, Jan 2, 2005) |
Summary
Over the past years, most of us have gotten so used to Intel's Pentium4 and a few selected Itanum2 that it has becomes kind of easy to disregard the rest of Intel's CPU repertoire, from X-scale to Banias and Dothan. Especially the Banias-core-based Pentium M, not to be mistaken for the Pentium 4M has earned itself the reputation of a superior performer at extremely low power. This lineage has continued to spawn the latest version, codename Dothan, featuring a 2MB Level2 cache that, in combination with the 32 kB each of instruction and data cache inherited from the Pentium3, makes for what could easily pass as Intel's finest. Only, it does not fit into the roadmaps heralding GHz along with superior heating performance.
Currently, no official desktop support for the Pentium M (as in mobile exists), yet, this does not preclude some mainboard manufacturers from taking the mobile 855 chipset with all its known inadequacies and transplant it onto an ATX form factor desktop board. The probably best known solution falling into this category is DFI's 855GME-MGF mainboard in ATX format and with a Socket479 processor interface. Supporting a meager DDR333 and AGP4X only and connecting the peripherals through a shared 266 MB/sec Hub-Interface 1.5 to the core logic, the specs are plebejan at best.
Specs very often do not do justice to the idiosyncrasies of any system, it is the symbiotic relationship between components that can overthrow the current grain of wisdom - especially if the latter originates in roadmap planning committees. Along came a Dothan ...
It is rare to write a report that could be involved in upsetting the PC world as we know it. For years, the Pentium4 family has been a more or less dominating force in the world of PC desktops and entry-level servers. For years, the P4 family has generated controversies about the relevance of high clock rates, which, in turn spawned terms like IPC and other performance ratings. For years, we have seen the development towards higher memory bandwidth enabled by so-called dual channel designs paving the way to DDR2 to satisfy the hunger for bandwidth of the double-pumped integer units. For years, we have witnessed the transitions from feeding a bit of extra power via the newly created ATX 2.03 compliant 4 pin 12 V connector to the CPU towards a highly sophisticated CPU voltage regulator module, totally uncoupled from the rest of the mainboard power. Finally, for years, we have watched dies shrink and coolers grow at about the same rate until both were hitting some unprecedented limitations in the form of current leakage and insufficiencies of the internal thermal transfers from the CPU to the integrated heatspreaders and finally to the heatsink-fan assembly.
Socket 479 (PentiumM) vs. Socket 478 (Pentium4). At closer inspection it turns out that the term socket 479 is actually a misnomer since the pin count is the same as in the case of the Pentium 4, only the pin B2 has moved into the A2 position. The Intel data sheet actually lists this pin as present on the processor package with a "reserved" function, which is the reason for the Socket 479 having that extra hole in the B2 position even though the pin is not physically implemented on the Pentium M in its currently shipping form.
If we believe the laws of marketing, the development outlined above was inevitable, we have Moore’s law and suffice it to say that the one undisputedly objective measurement of the quality of a CPU is, after all, still its clock speed. Rumor also has it that if something is repeated often enough, it finally becomes truth, courtesy of the principle of pluralism.
The competition, a.k.a. AMD has not slept during this time and the battle for supremacy in CPU-land has seen both sides taking the lead and losing it until the next processor stepping finally regained the crown. All of this is quite well established and part of rounding up the usual performance suspects. Behind the scenes, other battles have raged, optimizations of benchmarks for the one or the other architecture with higher geometry counts to favor floating point units or higher levels of texturing to take advantage of better optimized integer units including bumpmapping to generate geometry substitutes and fake FPU performance. The forces pulling in either direction were somewhat balanced even though everybody involved naturally always claimed that the grass was greener on the other side.
This carefully manicured equilibrium of ying and yang with its little quibbles over who is the fastest is suddenly usurped by a rather unexpected intruder in the form of a mobile processor: enter the PentiumM, codename Dothan.
|
Pentium M 735 (Dothan 1.7 GHz) |
next page: => Mobile Computing For Desktop Systems =>
If you enjoyed reading this article and found it useful, please consider making a small donation to LostCircuits.