|
LAME v3.91 MMX is what we use here to encode nearly 700MB of .WAV files as quickly as possible. It's almost entirely CPU bound so the stronger CPU will win here. It's interesting to note that this common version LAME isn't optimised for the P4 (with SSE2 and other optimisations) unlike many versions of media processing applications.
With the heavy workload we ask the CPU's to do here, an approximate 13% margin between XP2200+ and 2.4B shows just what a strong CPU Athlon XP is under certain conditions. It's worth bearing in mind that when evaluating CPU performance and trying to evaluate a winner, it's often pointless to do that in the general case. Each CPU has it's strengths and weaknesses and it's up to you the consumer to evaluate what you need from a CPU and purchase accordingly. An informed choice when buying in the high end can save you a lot of time and also money.
The seesaw nature of the benchmark results so far between each CPU prove that and also vindicate the choice of a good platform choice for each CPU. It also lets us highlight the weaknesses of each solution in each area too.
For our final look at the CPU performance we have SETI. Members of the OcUK forums have created a single test work unit from the popular SETI application. With this constant work unit, we can use it to evaluate CPU and system performance. Mentioned earlier, SETI is hard on the CPU and memory bandwidth. Can XP2200+'s FPU keep it close to 2.4B and its bandwidth advantage again?
For all intents and purposes, both platforms and CPU's are identical in SETI performance. While XP2200+ loses here, there will be units in SETI where it will be quicker than 2.4B and the chosen memory subsystem. It's worth noting that using RAMBUS RDRAM memory here with the 2.4B would have opened up the gap further. There is only so much that raw CPU performance can do for you before other system bottlenecks hold you back. XP2200+ is screaming for more bandwidth!
Lets go over what we've seen performance wise and try and come to some conclusions.
|
