swarm-support
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Ok, what about benchmarks


From: Darold Higa
Subject: RE: Ok, what about benchmarks
Date: Wed, 29 Jan 2003 12:41:47 -0800

Sorry if it seems dumb, I only did this side by side comparison when the guy I 
was working with asked how long a simulation run took for me and I told him a 
few minutes and he said that it was taking him hours.

Sorry for wasting everyone's time.

Darold Higa

-----Original Message-----
From: address@hidden
[mailto:address@hidden Behalf Of Marcus G. Daniels
Sent: Wednesday, January 29, 2003 12:20 PM
To: address@hidden
Subject: Re: Ok, what about benchmarks


Darold Higa wrote:

>OK, with all of this being thrown around about the different environments and 
>languages, has anyone done an informal performance benchmark?
>
>I ran my CasinoWorld simulation (the one from Luna and Perrone) using Swarm 
>1.4/Objective-C and got about 10x performance going from Windows 98 to Windows 
>NT/2000 and then another 10x Windows NT to Redhat (on roughly the same 
>hardware).  This fact prompted my co-author Michael Harrington, to move from a 
>Win2k environment to Redhat.
>
In general, Pentium 4s are the fastest by a significant margin -- see 
http://www.spec.org if you want to be convinced. I think the extra time 
Cygwin takes to set up is matched to some extent by the extra time it 
takes to resolve shared library symbols on Linux (or Solaris).  Once the 
executables are up and running, Swarm itself will probably be running a 
bit faster under Windows, since Swarm doesn't need to be position 
independent code.   However, this is compensated on Linux by faster 
system calls (Cygwin has to emulate some Unix semantics).
I don't think the get-more-memory system call speed is that much 
different.  The user-side malloc implementation used by Cygwin is mostly 
like the one on Linux (in glibc).  So I wouldn't count memory management 
in the `system call' Cygwin slowness category.

I think it is pretty dumb to make decisions about what system to use on 
these minor performance differences.  The big performance wins come when 
you start coming up with answers to the questions the profiler raises.


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.



                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]