Re: Parallel CM problem?

Matthias Blume wrote:

> In the meantime, I should warn that the parallel make facility does
> not seem to pay for its own weight in most cases.  In particular, I
> think that using more than 2 servers on SMP machines are hardly worth
> the trouble. 


Do you mean more than two servers period, or just more servers than CPUs 
  (or cores in this case)?  I'm currently building on a MacPro with two 
dual core processors.


> Spreading things across networked machines almost always
> loses more than it wins.  


Yes, I kind of expected that latency would pretty much outweigh any 
benefits.  At least for source files of a sane size.


> The whole thing has been very low priority for
> me, since nobody seemed to be using it.  You are the first to complain
> in earnest.  To be honest, my own testing had been confined to
> the example of recompiling SML/NJ -- where, by definition, there is no
> issue with the dynamic environment because the bootstrap compiler
> uses a separate link step.  So I never noticed the existing problem.

Okay, thanks for investigating this.  At least I know that it merely a 
result of inept configuration on my part :-)


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642