[Valgrind-developers] Core 2 'make perf'

[Julian S. <js...@ac...>]

Hi.  If anyone out there has a Core 2 machine I would be interested
in seeing the resulting numbers from 'make perf' in either a 3.2.1
or svn trunk tree, preferably from a run when the machine was
otherwise quiet.  Also the same for P4 Northwood and P4 Prescott
would be interesting.

I'm mostly interested in 32-bit mode numbers at this point.  Of
course amd64/ppc performance is important too, but what I'm interested
in is gathering data for/against the theory that the memcheck's 
performance is determined more by L2 cache size than by any other
microarchitectural factor (eg, how good your branch predictor is).

Thanks,

J