Intel Corporation is likely to pour some additional performance into i925X chipset – the company’s premier core-logic for high-end desktops this year. Some reports suggest that the company may pump up the maximum supported processor system bus speed and also add support for higher-speed DDR2 memory.
Intel i925X chipset formerly known as Alderwood is expected to sport 800MHz Quad Pumped Bus for LGA775 processors, dual-channel DDR2 memory at up to 533MHz, PCI Express x16 port for graphics cards and a rich arsenal of I/O capabilities, including multitude of Serial ATA-150 ports with RAID support and even software WLAN access point. VR-Zone web-site reported today that i925X not only could support DDR2 533MHz but also add DDR2 at 667MHz to the spec list.
Some other unofficial sources also suggested that Intel was considering whether to add or not support for Pentium 4 “Prescott” processors with 1066MHz processor system bus into the i925X chipset. They also alluded that appropriate Prescott chips in LGA775 form-factor are in the works.
Higher frequency of PSB as well as higher bandwidth or RAM typically give some additional performance boost for new platforms. 1066MHz Quad Pumped Bus will take full advantage of phenomenal 8.5GB/s bandwidth provided by dual-channel DDR2 SDRAM memory, giving end-users new height in performance computing.
Right now the Santa Clara, California-based company supplies Intel Pentium 4 Extreme Edition processors with 2MB L2 cache for gamers and performance enthusiasts. The Extreme Edition chips are substantially more powerful than typical Pentium 4 CPUs. The company said it would supply Extreme Edition processors for future systems that use Socket T. Current Pentium 4 Extreme Edition chips use Socket 478 and 800MHz Quad Pumped Bus. It is not clear how performance of Pentium 4 “Prescott” processors with 1066MHz PSB will compare to that of Pentium 4 Extreme Edition chips with 800MHz PSB.
Alderwood will also have some limitations in overclocking, VR-Zone notes. The clock speed or PSB will be locked down when it exceeds 12% increase over its default speed and the method is by tracking down the source clock. Mainboard makers are looking forward to overcome the limitation. The claim may contradict with the report about increasing maximal processor system bus speed on i925X.
Officials for Intel did not comment on the story.
Comments currently: 4
Discussion started: 04/13/04 08:25:04 PM
Latest comment: 04/14/04 06:34:45 PM
[1-4]
1.
They have heat issues to deal with, and no sign yet that they have managed that. If they limit to 12% overclock on their "enthusiast" board they are insane. If that turns out to be true please don't bother reviewing the damn things. And if the mfg's have to try to overcome that limitation you are talking increased prices on the mobo, and serious potential stability and compatibility issues now and down the line should you try to stick a Tejas in that 775 socket. Intel seems just full of bad news lately.
I have many products from them and I sure hope they get it together. The DDR2 thing is a big pill for anyone to get over in terms of cost, the heat and 12% overclock issues are going to kill their entire cpu line from enthusiast consideration, imo.
Of course this might be because those same enthusiasts didn't overwhelmingly pony up the $$$ for the Extreme P4's, go figure. Anyway, I guess AMD must be looking at this and thinking, ok time to build more FX chips! Intel I sure hope we aren't going to have to start writing "where the fall began at Intel" articles over the next year or two...
[Posted by: Anemone | Date: 04/13/04 08:25:04 PM]
2.
Well you all could say Intel is starting to get "bad", but where is amd?
If they had made AMD64 with 512kb cache from the start and dual channel for every one with "regular" memory, we could say intel is looking bad, but it isnt the case. AMD is not looking good to. At least for me.
[Posted by: I | Date: 04/14/04 03:20:28 AM]
3.
L3 cache, not l2 cache.
[Posted by: SOmeone | Date: 04/14/04 05:33:50 PM]
4.
I agree the dual channel thing is a bit sad, but in many cases they don't need the second channel. And yes while it's a stretch for the FX the chip performs well for that stretch. I think AMD has done admirably in bringing new tech to the market, stretching the limits of our current desktop abilities while not socking it to your checkbook. Intel just raises the $/performance and they just don't care to have a limit. The .09 process was "according to old marketing" supposed to bring a new level of cost to the chips, making them potentially far more affordable. Its not like you see that. I think AMD is giving you a lot more for the cost its costing them to make the chips than Intel, and then more performance per pure purchase price $ on top of that.
I think Intel is sinking into the corporate market and I think that when they can no longer compete for top performance they are just giving up and charging whoever is crazy enough to pay their prices. They'll have buyers but I'm betting they lose 10% market share over the next 6 months alone. Then they'll lay off saying they aren't making enough money when their costs have gone down, and their prices have remained the same.
I just don't think they care. I don't think they perform for the prices they are charging and I doubt very much they think the market will react at all to this.
I think they are wrong. My money at least, is going to the other side, after over a half dozen machines built on their platform.
$.02
[Posted by: Anemone | Date: 04/14/04 06:34:45 PM]
[1-4]