pete at he.iki.fi
Sat Jul 12 12:51:28 PDT 2003
> It's not as simple as that. In practice a lot of factors about
> your system and the type of work you're doing will affect the
> performance. On paper, the SA/KSE method is supposed to combine
> the best aspects of 1:1 (libthr) and N:1 (libc_r), and should
> threoretically be "better" than either one. But, in practice,
> complexity and overhead may drown out the performance gains.
> Conversely, context switching overhead may not be as great a
> penalty for the 1:1 model on modern cpus.
Anyone have any numbers of different architechtures for the context
switch? AMD64? Itanium? i386? PPC? Sparc? How about interrupt
latency? SMP coherency overhead?
The basic memory bandwidth has become a marketing thing so
the raw megabytes per second are readily available but in multiprocessor
and multithreaded environment there is a lot more into it.
More information about the freebsd-performance