[Shootout-list] Re: Integer overflow
Bengt Kleberg
bengt.kleberg@ericsson.com
Mon, 27 Sep 2004 18:50:32 +0200
Isaac Gouy wrote:
>>where does the switch bit enter into the test if we create lots of
>>threads (in a way that is different from how only creating 1000
>>threads is tested)?
>
> Last time I looked the test repeatedly sent messages between the
> threads.
sorry, but that is the message sending test. in it i do not want to
create lots of threads dependent upon n. instead i am advocating keeping
the number of threads constant (say 10000, or similar) and let n decided
the number of times the message goes around the ring. however, sofar i
have not been able to persuade mr fulgham.
i think you objected to this fixed number of processes since it would
give the languages that allocated a static array an unfair advandage.
correct? or am i unfairly accusing you of something you have not done?
anyway, i asked why this would be unfair since it is possible to
allocate an array depending upon n, if the number of processes changed
with n. perhaps i have missed the answer to this question of mine? have
you seen one?
no, this thread creation test i am writing about is another test i have
submitted to the shotout (called procinst). here n is directly related
to the number of threads. it has not been published/made official in the
shootout.
...deleted
>
> Even people uninterested in Concurrency Programming can be interested
> in thread creation times.
you are correct. but a few threads with clumsy interaction mechanisms
will just make the programmer in question think that threads is a bad
idea. there are lots of stories available where frustrated programmers
have used pthreads and come away thinking that threads is just a
disaster waiting to happen.
i prefer to try and give an example of how a good concurrency based
system would look.
bengt