To: Ali Chen who wrote (44056 ) 12/26/1998 6:44:00 PM From: Tenchusatsu Read Replies (3) | Respond to of 1573006
<Could be. At least their pick of K6-based systems for "mainstream" category is a real bottom - integrated video with apparently UMA frame buffer, and the slowest Quantum Bigfoots - It makes me real suspicious.> OK, you bring up a good point. Let's put aside our flamethrowers for a while and address this very real problem on AMD's hands. We all know that in a capable system, the AMD CPU's will shine and perform admirably. But so far, PC Magazine hasn't found too many name-brand AMD systems out there with choice components. If it looks too much like a conspiracy, feel free to e-mail those guys and suggest to them some AMD models that do feature the best components. The problem is that AMD still has the reputation of being a low-tier contender. Those system builders who are choosing AMD are choosing it purely because of price. Therefore, the other components that are going into an AMD-based system are also being chosen with regard to price, not quality. That's why quality components are more likely to appear in Intel systems over AMD's. If the builder is going with more expensive components, they might as well go with a more expensive CPU like Intel's, even though you can build a comparable system with an AMD K6-2 and a Super 7 motherboard. However, Anandtech's tests of the K6-3 signify a real chance for AMD to get rid of its low-tier reputation. Yet I am still concerned over the gaming benchmarks using the Riva TNT AGP card:206.132.42.114 Those benchmarks show the K6-3 450 behind a Celeron 300A in GLQuake 2, and behind a Pentium II 400 in Half-Life. You can argue that the 3D-Now drivers aren't there yet, that Super 7 chipset makers need to improve upon their AGP implementation, that Anand's tests are invalid, etc. But this is real bad news nonetheless because it shows the lack of timely 3D-Now and Super 7 support for games in AMD systems. It's as if AMD has taken a "Build it and they will come" approach to 3D-Now, but so far, not everyone is coming. So in conclusion, as we argue about the theoretical cases, "superiority" of architectures, and anything else that is worthy of an advocacy newsgroup, AMD will still have a problem with platform execution. I don't care if this is an Intel-backed conspiracy or not. The source of this problem is irrelevant. Results are what count, and I want to see real results from AMD-based systems, not theoretical comparisons from hardware junkies. Once I see that, I'll be long on AMD. Until then, AMD will still be a day-trading stock for me. Tenchusatsu