Chip makers Intel and Advanced Micro Devices are preparing their next round of server processors.
Intel is expected to launch its upgraded Bensley server platform next month at its Intel Developer Forum. Meanwhile, AMD offered a snapshot of a forthcoming Opteron chip, due later this year, at the San Francisco International Solid State Circuits Conference.
The new chips, which will all come to market in machines near the mid-year mark, are designed to raise the performance bar for the servers that are the backbone of corporate data centers.
The chips also serve to continue competition between the two chip makers. The x86 server arena has seen intense competition of late as machines containing AMDs Opteron chip have gained acceptance among businesses, and Intel has moved to counter by upgrading its server processors and rolling out new platforms to support them.
Intel will get the next round started by introducing the Bensley platform which promises a higher-performing processor and greater memory bandwidth.
A successor to its Lindenhurst platform, Bensley will pair Intels new 65-nanometer, dual-core Dempsey processor and Blackford chip set.
Blackford doubles up on busses and memory channels. It includes twin busses for feeding data into its two processors and four channels for buffered memory modules.
The buffered modules also work to lighten the load on the memory bus, thus aiding performance.
Intel will also include its virtualization, input-output acceleration and management technologies with the platform.
The chip maker is expected introduce the platform at its forum, where it will state the silicon is ready for prime time.
Servers from brand-name manufacturers who are expected to adopt Bensley—companies such as Hewlett-Packard—may not arrive for another 60 to 90 days, however. One person familiar with the matter said to expect them in May.
An Intel spokesperson said the company is on track to ship the Bensley platform in the first quarter. However, he declined to comment on the companys plans.
Intel will also likely discuss Woodcrest, an all new follow-on to Dempsey, at the forum.
Woodcrest chips, which Intel has said will consume considerably less power—about 80 watts, which is significantly lower than existing Xeons—will plug into Bensley platform servers and come out in the second half of the year, the chip maker has said.
New Line of Opterons
AMD, working to maintain its momentum, will roll out a new line of Opterons at mid-year.
The company, at the ISSCC, discussed one of those chips, a dual-core Opteron it said would run at 2.6GHz and offer double the memory bandwidth by incorporating a DDR2 memory controller.
The new controller, and the subsequent move to speedier memory, will allow AMD to bump up server performance without necessarily increasing its chips clock speed or its power consumption, it said in the presentation.
The new DDR2-equipped Opterons are expected to appear in different configurations—AMD offers different Opterons for one, dual and multi-processor servers for example—and to use a new socket, in addition to including AMD Virtualization Technology.
The 2.6GHz Opteron discussed at the show will consume 95 watts of power, AMD said.
However, the company indicated in its presentation that it could offer lower power versions running at slightly lower clock speeds as well. AMD is likely to do so as well.
Whereas server makers are generally slower to adopt new technologies, given their critical roles, than PCs, the new chips impacts are likely to be stretched over time.
However, “Its safe to say that obviously both companies are trying to make revisions to their product offerings that keep them competitive,” said Dean McCarron, principal analyst at Mercury Research.
“The market obviously has changed from a few years ago. Now that we have AMD present as a stronger player, its going to be very interesting.”
By taking a different approach and choosing to build its memory controller into its chips, whereas AMD used an off-chip memory controller hub connected by a bus, AMD bought it self an edge in performance, McCarron said.
AMD has seen share of x86 server chips grow over the past several quarters. During the fourth quarter of 2005, its server share increased to 16.4 percent from 12.7 percent in the third quarter of 2005, figures from Mercury Research show.
Overall, AMD gained 3.7 market share points to 21.4 percent during the fourth quarter, according to Mercury Research.
Intel, which has historically garnered 80 percent or more of the x86 chip market, saw its share fall 3.8 points to 76.9 percent, Mercury Researchs numbers show.
“AMD gained share in every segment. Serer contributed, but so did desktop and so did mobile,” said Dean McCarron, principal analyst at Mercury Research.
Although it may eventually use FB-DIMMS—it hasnt ruled out using them—AMD chose to stick with standard DDR2 for the moment.
“They are taking fairly different approaches. The market will tell us which one is right,” McCarron said.
Meanwhile, Intels Dempsey chip may well have a short stay as server makers adopting the Bensley platform may move rapidly to Woodcrest when that chip comes out later in 2006. Server makers have done so often in the past, McCarron said.
Intel will also beef up its four-processor and higher systems.
For that market, is planning to expand its Xeon MP line with Tulsa, a dual-core Xeon MP processor that packs more than 18MB of cache—including 16MB of level 3 cache and 1MB of level 2 cache for each processor core—during the second half.
The chip maker disclosed this week at the ISSCC that Tulsa will pack more than 1.3 billion transistors and measure 435 square millimeters.
Running at 3.4GHz, it will have a thermal design power of 150 watts. Although thats about 15 watts less than the companys current Xeon 7000 series chips, whose thermal design power specification is 165 watts.
Aside from minting Tulsa on its 65-nanometer process, which generally helps reduce power consumption, Intel has taken measures to lower the caches power demands as well as to keep it fed.
The chip will employ a controller that connects its two processor cores and its cache to its external bus.
Intel will use Tulsa to update its Truland platform, the company has said.
Editors Note: Jeffrey Burt contributed to this story.