[hashcash] Re: Obtaining speed results for different CPUs (collision tests/sec)

  • From: Jonathan Morton <chromi@xxxxxxxxxxxxxxxxxxxxx>
  • To: hashcash@xxxxxxxxxxxxx
  • Date: Wed, 25 Aug 2004 00:31:22 +0100

Thanks very much for that Jonathan, that's a great help. I had a result for a P2 400MHz of about 210,500 (hashcash v0.28, which I don't think uses fastmint either). So using your method of extrapolating down from that, works out at around 6,500 for the 486/25 (although the P2 architecture is different again to the MMX)

The P6 core (which the P2 is built on) is very very different from both the P5 and 486, so the extrapolation is no longer valid for that. In fact, it's rarely possible to compare two different generations like that - the P5/486 is an exceptional case.


I'm going to take these upper (G5) and lower bounds and start plugging into a few breakeven estimates.... Although I think I'll multiply the 486 result by 4, to pretend that we had fastmint code working on it.

No, multiply it by two instead. A 2x improvement under fastmint is typical for scalar-only machines, such as the 486 and pre-G4 PowerPCs. The 4x improvement on the Pentium-MMX is due to a further near doubling under the MMX optimisations, which aren't available on the 486.


--------------------------------------------------------------
from:     Jonathan "Chromatix" Morton
mail:     chromi@xxxxxxxxxxxxxxxxxxxxx
website:  http://www.chromatix.uklinux.net/
tagline:  The key to knowledge is not to rely on people to teach you it.


Other related posts: