Does anyone have some expiriance with CPU architecture related MT performance?
Does MT performs better on AMD Sempron64 that has:
- on-die memory controler and lower memory latency
- shorter pipeline and less penalty for pipeline flushes
of Celeron D that has:
- higher clock
- larger L2 cache
I’ve seen that Intel’s Netburst architecture is better performer in predictible aplications (like Pi calculating, media encoding) and AMD seems to be better in aplications with a lot of unpredictable branches (like games).
The only thing I don’t know is:
In which of those two aplication groups does MT belong?
Any real-life expiriance?
