Author: Matt Taylor
Date: 15:01:48 12/02/02
Go up one level in this thread
>>>I hope they can deliver a quad opteron for a resonable price. They were talking >>>about quad >>>K7's two years ago and not a single instance has shown up yet. Intel talked >>>about the 8-way >>>boxes a while back and delivered a kludge there, using a "fusion" chipset to tie >>>two 4-way >>>clusters of processors together into a single 8-way box, but with terrible >>>memory performance. >>>They tried to offset that by only offering 2M L2 caches, but that drove the >>>price up and didn't >>>help memory-bound large applications at all... I hope the quad opterons don't >>>end up in >>>never-never land as the 8-way boxes did.. >> >>Here's a picture of a Quad opteron system if for some reason you think it's >>never going to happen... >>http://www.amdzone.com/articleimages/cpu/hammer/4popt.JPG >>There are many Dual Opterons out as well.. > >They had "pictures" of quad K7 MBs as well. Never saw one on the street, >however. >Again, I don't see how to evaluate what's gonna be. Just what is that we can >get our hands >on... Actually Opteron scales to 8 processors and should scale up further. Opteron is a ccNUMA architecture unlike any current x86 systems. (If you look around, you -CAN- find ccNUMA systems built from dual-Xeon and quad-Xeon nodes.) >>>If I recall, the 4=way dual 2.0ghz xeon is the fastest PC-class machine around >>>right now, >>>by a wide margin. And the heavier the load placed on it, the wider that gap >>>becomes... That depends on what you're doing, who you ask, and how you optimize. One can summarize by saying that Athlon will run existing code -much- faster than a P4. Hand-tweaked programs can run faster on P4, and memory-intensive algorithms employing SSE will run faster on P4. On a P4, code can be more lax with stack usage because the P4 has a lower L1 latency (2 clocks) than Athlon (3 clocks). On Athlon, a shift (1 clock) is less costly than on a P4 (4-6 clocks). Keep in mind that using the AGU uses the shifter on the P4 because the AGU circuit was removed, too. Many other differences exist, but there are many well-written articles documenting them, though I don't think I've seen any on Tom's Hardware. A careful observer will also note that code that runs efficiently on a P4 will likely make inefficient use of the Athlon and vice versa. It's unfortunate that benchmarks do not include source. Is this the principle of cold fusion research?
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.