[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Print to SHR
Andy McFadden wrote:
aiiadict@gmail.com wrote:
On Mar 6, 12:28 pm, "Michael J. Mahon" <mjma...@aol.com> wrote:
While Moore's "Law" has pretty much petered out for improvements
of single-thread processor performance,
When did the 2.5Ghz come out? It seems like there have been ZERO
speed improvements in the last ?5? years.
The clock rates have stabilized or retreated. My Intel QX6700 is clocked
at 2.66GHz, slower than the three-year-old 3GHz Pentium 4 it replaced.
However, various other improvements made it *significantly* faster for
single-threaded integer performance (as determined by playing old games
in DOSBox).
Yes, in the interest of power saving, clock speeds have dropped,
but the procesor microarchitecture has become more efficient, resulting
in a small improvement in throughput.
I don't think it's doubling every year though.
I sure isn't--or even every two years.
Since they have run out of gas to go faster, they've now shifted
the pitch to "multicore" machines, that is, shared-cache symmetric
multiprocessing.
I'm a big believer in parallelism, but the world hasn't figured out
how to use parallelism either efficiently or reliably. Threads are
the worst possible model for parallel processing, since they present
all of memory as shared by all threads!
In another 20 years or so, graduates can be expected to be reasonably
competent at parallel programming--until then, good luck. How well
the industry survives the performance and reliability slump for that
long is anyone's guess.
-michael
NadaNet networking for Apple II computers!
Home page: http://members.aol.com/MJMahon/
"The wastebasket is our most important design
tool--and it's seriously underused."