Kenn
Please, just shoot me now. I've heard all the hype surrounding the K7, and while it certainly looks good, it's not the first time an Intel-killer has been announced. Look at the original K6. Back before release, it was slated to be a powerhouse competitor to the Pentium Pro, but by the time it was released, Intel had enough time to push ahead with its Pentium II line, quickly making the K6 a non-factor.

Further back, IBM and Apple, so proud of their new PowerPC 601, laid out a roadmap showing quarterly doublings (even quadrupling) of performance. Exploiting even further the FUD (fear, uncertainty, doubt) principle, there was word of a 66Mhz 64-bit PCI bus developed independently at IBM. We all wondered then what would be powering the guts of our new computers, but again, there was Intel, steadily increasing the performance of their processors and releasing newer, faster designs on the "old, tired CISC platform."

Going back to the K7, let's take a look at the details. The L2 cache is going to be integrated into the CPU cartridge, running at 1/3 or ½ CPU speed. At 1/3 speed, that's slower than the 200Mhz bus. At ½ speed, the cache is faster than the system bus, making much of that speed increase irrelevant (look at the lackluster performance increase for Pentium II moving from 66mhz to a 100Mhz FSB). At 1/3 speed for 500Mhz, the cache is slower than the bus anyway, so what's the point? While the K7 may appear strong on paper (and indeed in tests), the momentum is with Intel, who's pushing the limits with the upcoming 133Mhz Camino chipset and Rambus RDRAM.

Far from being a panderer to every whim of Intel, I'm still going to hold my breath and take an agnostic stance on the K7.