what do you mean by pipe performance ?!
instruction pipelines are utilised to reduce the amount of cycles required to perform an instruction, AMD's SSE2 and 3 support in some cases take more cycles to complete than a similar clocked Intel processor of the same generation due to having longer pipes or shorter depth.
more cache however does reduce the chance of a stall in the execution pipeline.
aha , i understand now why intel performs better than AMD per clock , apart from better architechture for CPU circuits
!
learned something new !! thanks alot !!
L3 is slower (slightly) than L2 because of access time. You're better off in EVERY case having more L2 than L3. And as for "Raw processing power" that's literally based on your Lcache speed (your ghz :: cache relationship) so of course it makes a difference. In fact it probably makes the biggest difference.
Access times aren't the biggest of problems. by reducing the chance of execution stalls in the pipeline by reducing the chance of cache miss you can improve performance by not having the cpu performing short cycles.
(10-29-2010, 05:48 AM)Squall Leonhart Wrote: [ -> ]AMD chips are limited in instruction pipe performance for the most part
Now the last question will remain !!
- for someone with an AMD platform and using his PC for dolphin ... will he make use of extra cache L2 or L3 ?
what about north-bridge freq. <Hyper-transport> does it make any difference ??
hyper transport is the replacement for the FSB on amd chips. it mostly only matters when overclocking or using otherwise usually unsupported memory speeds.
(10-30-2010, 01:42 AM)Squall Leonhart Wrote: [ -> ]hyper transport is the replacement for the FSB on amd chips. it mostly only matters when overclocking or using otherwise usually unsupported memory speeds.
what about cache ?
is it useful to increase cache while using an AMD platform ?
i mean for dolphin of course
He already answered your question, yes. With modern AMD cpus cache size is even more important than with intel cpus.