Subj : Re: What is the real costs of LOCK on x86 multiprocesor machine?
To   : comp.programming.threads
From : chris noonan
Date : Fri Jul 29 2005 09:48 am


Mirek Fidler wrote:

> Measuring the same on my uniprocesor AMD64 machine, LOCKed instructin
> seem to be 3 times slower than regular one, but I guess that has only a
> little relevance on real MP machine.

Surprisingly, on Intel Pentium processors the overhead of
a locked instruction is about the same for a single processor
as for one in a multiple processor configuration.

The way Microsoft (for example) avoid the performance hit is
to replace the LOCK prefixes in uniprocessor builds of their
operating system binaries with NOP instructions.

Chris

.