Foresight (at the time) told you NetBurst was a dead end though. AMD made a lot of sales from people avoiding NetBurst, ignoring clock speeds and looking at act

Foresight (at the time) told you NetBurst was a dead end though. AMD made a lot of sales from people avoiding NetBurst, ignoring clock speeds and looking at actual performance numbers. NetBurst looked good on benchmarks; code with few/no branches that fit in the L2 cache if not the L1 cache. On real world code consisting almost entirely of branches and pointer chasing the NetBurst chips performed horribly compared to lower clocked Athlons. Intel's promises of 5+ GHz chips got more laughable as the line went on because they were already straining cooling and power supply options at less than 3GHz.