need fast parser for comma/space delimited numbers

[Rick Nooner]

On Mon, 20 Mar 2000 08:06:55 +0000, Tony McDonald
<tony.mcdonald at ncl.ac.uk> wrote:

>In article <etdem96rar0.fsf at w20-575-110.mit.edu>, Alex 
><alex at somewhere.round.here> wrote:
>
>> > what pystones are you getting on  your Sparc 10?
>> 
>> athena% ./src/Python-1.5.1/Lib/test/pystone.py
>> Pystone(1.1) time for 10000 passes = 3.06
>> This machine benchmarks at 3267.97 pystones/second
>> 
>> I have never used pystone before, so please let me know if this is the
>> right way.  Noone has bothered to upgrade python for a while around
>> here, but if you can't get the figures from anyone else, I can
>> temporarily install the latest python and run pystone on that, too.
>> 
>> Hope this helps.
>> 
>> Alex
>
>That seems the right way to use it Alex.
>Thanks for taking the time out to do that - I'm going to have a little 
>look around our systems here and do some snooping. Our most powerful Sun 
>box (Enterprise 450) if I remember rightly) gets 4500 pystones.
>
>I believe that there is a small performance gain in using 1.5.2 so 
>perhaps you ought to upgrade?
>
>thanks again,
>tone.

Here are some figures from some of my machines:

	Netra T1120 	2x296 Mhz CPUs	3496.5 in 2.86 secs
	Enterprise 250	2x296 Mhz CPUs	3558.72 in 2.81 secs
	Enterprise 450	4x296 Mhz CPUs	3558.72 in 2.81 secs
	Enterprise 5000	6x248 Mhz CPUs	2695.42 in 3.71 secs

This shows very clearly that pystone.py is only measuring a single CPU
which is expected since it is a single threaded, single process
program.

Rick Nooner