[hashcash] Re: Low end benchmarks
- From: Jonathan Morton <chromi@xxxxxxxxxxxxxxxxxxxxx>
- To: hashcash@xxxxxxxxxxxxx
- Date: Sat, 28 Aug 2004 20:09:27 +0100
Following on from the recent posts about minting speed on a 486, here
is
the benchmark output from a Cyrix 5x86 100MHz. Not quite the same
thing,
but back when I had access to a 486/100MHz as well they were almost
identical in performance so this should be a fair guide:
Rate Name (* machine default)
32266 ANSI Compact 1-pipe *
28576 ANSI Standard 1-pipe
25980 ANSI Compact 2-pipe
4606 ANSI Standard 2-pipe
--- PowerPC Altivec Standard 1x4-pipe (Not available on this
machine)
Best minter: ANSI Compact 1-pipe (32266 hashes/sec)
I was amazed by the dramatic difference in speed of the Standard
2-pipe from
the others, but I repeated the test two more times and the results were
consistent.
It's probably because the Standard 2-pipe core has the largest code
size of the above set (you're using an old version, BTW), and it
therefore overflows the 5x86's cache. Instruction-cache misses tend to
be even more lethal to performance than data-cache misses. A similar
result is likely on a real 486, which I believe has an 8K+8K Harvard
cache.
I'm in the middle of installing Linux on a couple of older machines -
including a 25MHz 486DX - so real tests of this nature will be easier
to come by. Of course, because I'm installing Gentoo (to be consistent
with my other Linux machines), this is going to take a while. (The
other machine is a PowerBook 5300ce which I've disassembled and plan to
turn into a "living artwork". Not directly relevant to hashcash, but
it will in fact be easier to run hashcash on it when Linux is running
properly.)
--------------------------------------------------------------
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: