Multi-core processors

Erik Wikström erik-wikstrom at telia.com
Fri Apr 15 12:45:14 PDT 2005


Danial Thom wrote:
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2388

the problem with technical analysis is that a
case can be made for or against any product
depending on what tests you do, and most sites,
like most techies, are anti-intel-biased. The
comment at the end of the article is a perfect
example: they blantently ignore the possibility
that you might, just might, be doing 2 or more
things at once on your desktop. With multi-core,
the benchmarks have to be multi-tasked not just
multi-threaded. I don't only want to know how
fast application A is on single and dual core,
but also how fast it is when application B is
also hammering away. 
I choose the article because they actually did some
multitasking tests where a noticable difference could
be seen. However one should keep in mind while reading
the article that they are mostly interested in tests
which represents the usage of a normal Windows or
gamer. And for those people gaming performance is what
counts, and games does not yet benefit from multicore.
For servers things are different and a good admin will
know if multicore will be beneficial for him or not.
However multicore is the future right now so more and
more applications will be multithreaded which means that
multicore will be beneficial for almost everyone soon,
but maybe not just this year.
--
Erik Wikström




More information about the Users mailing list