Formerly About Advanced Micro Devices Message Board

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Gopher Broke who wrote (94234)	2/19/2000 10:47:00 AM
From: Scumbria	Read Replies (1) of 1570917

Gopher, No doubt Willamette will run faster when the pipe is full, but the performance is so much more dependent upon successful branch prediction that I think we may see some strange anomalies introduced by the higher pipe latency. Compiler optimizations will play an even more important role in keeping the pipe primed, and how will pre-Willamette optimized code perform? A deeper pipe will create more bubbles on mispredict, but the duration of each bubble is shorter than in a shallow pipe. The net effect is that a deep pipe runs faster, and has only slightly longer absolute latency on misprediction. All empirical evidence points to the overall performance advantages of a deep pipe. Alpha held the integer performance lead for a decade, based on a deep pipe. Athlon has a deep pipe, PIII has a deep pipe. Willamette will likely be the highest performance integer CPU, as a result of the higher clock rate. The stack based architecture of x86 FPU is crippling, so it remains to be seen how the floating point performance will stack up against RISC machines. Scumbria

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next