Originally posted by Northbridge
Xeon processors can be better if a program can recognize them, but you can't say a computer with two 2.8GHz Xeons has a an overall 5.6GHz clock speed.
It wouldn't perform as well as a single 5.6GHz processor, but you would see a marginal increase in performance.
You don't understand the nature of SMP do you? The processors would still share a common I/O bus and memory subsystem. No matter how you look at it, the system resources are shared by MP. I own a dual system myself and I know it could outperform a high level p4. You have to remember even if a program wont use both cpu's the most important thing will, the operating system. So even if program "A" doesn't support SMP, if processor number 1 is busy it has processor 2 do the work. That's just a very simple example. In reality the operating system already has the resources divided, such as one chip will take care of threading while the other does math, or it will have one chip take care of certain process's critical to the OS while the other takes care of non-prioritized things. This is why an SMP system is better than any single chip. Their is a reason why it is called multi-processing. Just because an app cant use both, doesn't mean the OS wont manage both cpus to increase performance. Also, when you have two processors your get double the IPC. One P4 could never beat the combined IPC power of two P3's. A dual P3 system could easily take two or more requests at once, where a P4 would not be able to do nearly as well.
P3 can do 6 IPC
P4 can do 6 IPC
So 1 p4= 6 IPC
Two P3= 12 IPC
IPC= Instructions per cycle/clock.