JC heeft een paar Bryce3D benchmarks te pakken weten te krijgen. Check this (met een korreltje zout):
More leakage from one of the folks I know with one of those nice Athlonic K7 sample systems. It seems a full production release model, I would guess. This particular person reported to me that his PIII-550 completed the Bryce 3D benchmark in 123 seconds, whereas his K7 system completed the task in about 86 seconds. That's about 69.9% of the time, or about 43% faster. If somebody could mail me their scores with a PIII-550 so I can verify, I'd be delighted. As it stands, this would seem to make the 550MHz K7 more similar to a PIII running at about 800MHz, if the benchmark is linear with clock speed, and if the PIII number given isn't accidentally inflated by some unknown or unseen bottleneck. Mind you, the less linear Bryce is with MHz, the more that the K7 may have an advantage, since such benchmarks have less effect with MHz increase -- if this is the case, the PIII would have to be even higher clocked to compete. Anyway, somebody get me their PIII-550 results so I can more properly look this over. By the way, I'm trying to get my hands on some rc5 numbers for the K7 -- I'm sure nobody's really interested, but I thought you might like to know.