On Tue, Oct 16, 2007 at 08:48:09AM -0000, Michael Prokop wrote:
> I'm not really a friend of non-free software. What's the benefit of
> prim-net compared to something like cpuburn (which is shipped by
> grml already)?

The torture test of mprime stresses CPU, L1 and L2 caches at the same
time. It probably does not put as much load on the cooling system as
burncpu does but it is probably closer to any high workload real world
example.

It also compares the results of the torture test against known good
results, so any error in complex calculations should be detected.
It also has optimized code paths for cpus newer than P6/K7 which are
autodetected.

I did not find any real reasons why mprime should be superior to
burncpu, but it is used to test system stability. Also, having a choice
between these two programs would only benefit the user imho.

You can also get the sourcecode to mprime minus the network
communication layer part on http://mersenne.org/

    Ingo
-- 
"Niemand hat die Absicht eine Mauer zu errichten"
	-- Walter Ulbricht, DDR-Staatsratsvorsitzender
	   http://video.google.com/videoplay?docid=6233111545394191802