Replies to post #64370 on Advanced Micro Devices Inc (AMD)

[wbmw]

Re: Your argument is religious in nature, rather than scientific.

No, there is a basis of fact involved. I can look at the benchmarks of Dothan, which is Pentium M based and has plenty of quantitative analysis on it. I can then extrapolate the known specifications of Yonah (SSE3, FPU and SSE enhancements, low power enhancements, VT, etc), then Merom (slightly extended pipeline, additional issue width, increased buffers, EM64T, far greater memory and bus bandwidth, more low power enhancements, etc) and come up with the expectation that it should outperform K8 micro-architecture on a clock for clock basis while being competitively low power. I also expect it to clock much higher than today's Pentium M, given that the power envelope has been extended and it's made with desktop/server design parameters. In fact, it should clock above 3GHz on 65nm, while AMD will be stuck at 90nm this year (2.8GHz max for dual core, IMO). There's still a lot of unknowns and some faith that more will be involved than just a simple Pentium M derivative, but one thing is clear: it should definitely present a competitive product in the 2006-2007 time frame. AMD's K10 is an even bigger unknown at this point, but I will have to see how far Woodcrest outperforms Opteron later next year before speculating on whether AMD has a good chance with K10 on turning the tables.