On 4/18/07, Patrick McNamara <[EMAIL PROTECTED]> wrote:
http://www.theinquirer.net/default.aspx?article=38964

Indeed.  It appears that, at least temporarily, manufacturing
processes have outpaced engineer's ability to effectively use said
manufacturing capacity.  The absolutely massive multi-megabyte caches
on recent CPUs hint at the same problem.  There's long-term hope in
cores like Sun's UltraSparc T1, Via's C7, etc...  but ultimately, the
software _and_ underlying CPU architecture are going to have to handle
massive parallelism (one easy way to use lots of transistors) in a
reasonably transparent way, or use more transistors being fpga-like,
or a combination of the two.  Of course, as always, more and more
devices will fall into the silicon black hole that is the CPU as well.
Undoubtedly we'll eventually see high-performance machines that are
capable of functioning with just the CPU, with a few high-speed
interfaces for additional hardware or co-processors.

Wake me up when system RAM succumbs to the inevitable.

--tim
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to