Slashdot Mirror
Intel Pushes Back with Xeon 5100
conq writes "BusinessWeek has a piece on Intel's newest chip, the Xeon 5100, which many consider might be the chip that will llow them to stop losing ground to AMD. From the article: 'During the presentation, Intel ran the now-standard comparison test against AMD's highest performing chip, handily beating the system in a speed test. And in a jab at AMD execs, who handed kill-o-watt meters to analysts at the outfit's recent technology day, Intel execs used the same device to measure the new Xeon 5100 system's performance — gauged to be 7 watts better than that of the AMD-based system.'"
We are beginning to see why Apple made the jump to Intel.
It's not that they had anything that was all that much better than IBM or AMD at the time they were making their pitch to Jobs. It was the fact that their immediate future was being prepped with some impressive technology, both in terms of speed and speed-per-watt, which turned the Steve's head.
Information wants to be anthropomorphized.
They may be faster, and they may consume less power, but IT is still about cost. Something tells me, that with this great advancement comes a higher price tag than AMD.
Who cares is Intel is a few mips faster?
Intel has by far the largest fabrication capacity of any chipmaker in the world. Both IBM's and AMD's fab capacities are much lower (AMD has used IBM's fab to help meet demand). IBM's inability to produce high numbers and high yields led to the Intel switch. Remember the delay in introducing the iMac G5? Apple had the design ready, IBM couldn't produce the chips. Result: months go by without any iMacs to sell. More than anything technical reason, IBM was bad for Apple's bottom line.
How well does this new chip work in systems with more then 2 of them? How bad will the FSB bottleneck get when intel start using quad-cores with 2 duel cores linked by a FSB? What will socket F bring to AMD? When will we start to see 64 Bit tests? I think amd will be faster then intel in 64 bit mode When will open HyperTransport bring to the server market? What will intel do to beat HTX co-processors cards?
I would doubt that AMD chose to advertise specifically on this article. Most likely this is some adwords or related content scanning system which says "The article is about CPUs, we'll insert the ads of the advertiser who won the bid for that keyword"
I'll never make that mistake again, reading the experts' opinions. - Feynman
Since Japan has already hit Intel for anti-competetive moves, can AMD prove illegal dumping?
Which is why these 64-bit Linux benchmarks show that Woodcrest scales as good as (and sometimes better than) Opterons at 4p. The vast majority of x86 servers are in the 4p range. Even Opterons have a worse-than-expected scaling issue past 4p, anyway, if you bother to look around to find the benchmarks.
The Optron's scaling issues beyond 4P is not "worse then expected," because it is entirely expected of the architecture.
The high-end Opteron has 3 HT links. This means it can work with up to 8 sockets "gluelessly," but it really performans much better with 4-socket systems. The architecture for a 4-way Opteron server uses the extra HT link to reduice the number of hops, so only one case has two hops.
But you can imagine that the 8-way configurations have a much higher average number of hops between processors, PLUS much more data flowing over the same HT links. No, the K8 Opteron is not really designed well for 8-socket systems.
But K8L IS designed for 8-socket systems.
Take a look at a page on this in the K8L preview article on Real World Technologies. Adding a 4th HT link will really make a difference.
4-socket K8L systems benefit because they take advantage of the 4 HT links to provide 1-hop latency to all sockets in the mesh, and can now have external I/O hooked up to ALL processors.
8-socket K8L systems take advantage of two things: the extra HT link is beneficial, and the advanced mesh created by splitting up the HT bus widths means MUCH better performance for 8-way systems.
Woodcrest is impressive as hell, but I will tell you one thing: there's no way in hell it's going to scale well beyond 4-socket systems. This is for the same reasons that have been holding back performance on 4-way Xeon syetems (reduced bus speeds with 4 processors on the bus, too much traffic). The Dual-Independent Bus allows Intel to scale well to 4-way, but no higher. K8L will allow for glueless scaling to 8-way, and will still provide a a cheaper solution than Intel's Dual-Independent Bus for 4-way chipsets and motherboard designs.
Man is the animal that laughs.
And occasionally whores for Karma.
As long as Intel maintains this architecture, with a single data bus for RAM, PCI, PCI-e, AGP, BIOS, and other integrated functions, they'll be behind AMD. AMD's current (and future) HyperTransport provides a wider, more efficient data path than the front-side bus. AMD's per-processor memory controller scales past two sockets in a way that Intel just can't match. By pushing fully-buffered DIMM's, each with its own memory controller, Intel is ceding the design point to AMD: a single memory controller is too much of a bottleneck, the load needs to be spread around. This is especially true when you go beyond two processors in a machine, but even dual-socket boxes benefit from distributed memory controllers. Sure, the Bensley FSB goes to 1033Mhz from 800Mhz, but that doesn't sound like a big jump.
Until Intel has a real answer to HyperTransport, they'll be losing the high-performance, 4+ sockets market to AMD. For smaller two-socket servers, Intel will have to pay the RAM and/or server vendors to make FB-DIMM's price competitive with different flavors of DDR.
The other neat thing about FB-DIMM, from what I've heard, is the socket changes in AMD models every time they implement a new memory technology will no longer be necessary because the FB-DIMM pinouts don't change.
You are still innocent until proven guilty. What's changed is what they do to innocent people. - notnAP, #26891325