| This makes much sense to me as well, though I can't quite explain why... every "thought experiment" I come up with contradicts that theory... The P4 core traditionally does a better job of masking memory latency by sheer pipeline length, while the Athlon core tends to read memory in short bursts, which make their way through the pipeline more quickly - lower latency should be of greater benefit. Yet, the evidence stands...
__________________
|