Replies to post #61240 on Advanced Micro Devices Inc (AMD)

[wbmw]

Re: The only unknown is how good the AMD64 implementation is compared to the original because Prescott's was up and down in performance in that respect.

I don't know why this gets sensationalized so much. Performance tests indicate that Intel's 64-bit implementation is better than AMD's in some respects and worse in others. This should come as absolutely no surprise, since Intel's is based on their own micro-architecture, just tuned so that the op-codes are compatible with AMD's.

Merom's will be yet another micro-architecture, and the 14-stage pipeline will ensure very little reuse from Pentium 4. I would therefore expect Intel to have started from a clean slate and optimized the pipeline for high performance. It should be equal or better than the current implementation, most likely better.

Re: AMD told Anand there will be no new cpu architecture in 2006 so 65nm K8s look like the immediate opposition.

If they can get to 65nm. There has been nary a whisper about AMD's 65nm process since they showed a picture of an SRAM wafer at their last analyst's meeting. The rule of thumb is at least a year after tapeout. Has any 65nm AMD processor even taped out??

[Dan3]

edit duplicate

[Dan3]

Re: AMD told Anand there will be no new cpu architecture in 2006

Jerry Sanders and Dirk Meyers laid out AMD's current CPU architecture, AMD64, dual core, Hyptertransport, 14 stage pipeline, and on-die memory controller in 1998 and it hasn't changed since.

Seems to have been a real winner, too.

Maybe dumping everything, reversing course, and starting over from scratch every couple years (the way Intel does it) isn't... Ideal.

At least now they seem to have given up and are just copying whatever AMD does.

Mayber that will work out for them.