[Shootout-list] Re: Integer overflow, process creation

[Bengt Kleberg]

Brent Fulgham wrote:
 > I agree we should add this test, and will do so ASAP.
 > In case I lost it in my mailer mishap of a few weeks
 > ago, can you please forward it to me again?

more info about the process instantiation test. there are operating 
systems where creating a process and starting it are different steps. i 
want the created processes in this test to actually execute a few steps. 
then wait for something. i could mandate what, but i think it does not 
matter.

why keep them all around? if we take the ''create processes and delete 
them, repeat N times'' to the extream we get N times 1 process 
creation/deletion. this sounds as it would give a benefit to languages 
where processes are poorly impleneted (imho).
i am want to listen to arguments as to why it is better to have fewer 
processes created/deleted several times. but i do not think i will be 
persuaded by the argument that few proceses are easier for a language 
runtime. that goes against the spirit of benchmarking, as i see it :-)

now i could easily be persuaded to add a explicit deletion of all the 
created processes if that would make anybody happy. and, as a final 
peace offering: let us have 2 tests. one with N proceses, and one with 1 
create/delete N times. (i am beeing reasonable. i know i will regret 
this :-)


bengt