🚀 go-pugleaf

I just performed an regular batch compile of our project, which has a 
moderately large number of files,
on a dual processor system - while watching activity using the task manager.
It shows that _both_ processors are very active during the 'make'; there may 
be 'wasted' cycles, due to Windows XP
taking idle time, but it's not the case that one processor does nothing 
while the other is doing compiles.
In fact, they both have the same utilization curves, and peak at the same 
times.

Have you looked at your CPU usage during compiles, either with the IDE or 
using batch files?

Norm
"Jonathan Benedicto" <incorrect@no.server> wrote in message 
news:432f0452$1@newsgroups.borland.com...
> "Ed Mulroy" <dont_email_me@bitbuc.ket> wrote in message 
> news:432efebb$1@newsgroups.borland.com...
>> So have your program launch the compiler instead of launching make.exe.
>
> My app is actually run by make. When someone uses my app, they change the 
> line in the makefile "BCC32 = bcc32" to "BCC32 = mtbcc32". Then, make 
> calls my app "thinking" that it is bcc32, so my app gets a command line 
> like this: mtbcc32 CompilerOptions File1 File2 etc.
>
>> Now have your program launch bcc32 twice in this way:
>>
>> bcc32 +TheOpts.cfg @1stListFile.txt
>> bcc32 +TheOpts.cfg @2ndListFile.txt
>>
>> Each instance of bcc32.exe will compile all files listed in the 
>> redirected text file whose name it is given.
>
> The problem I have with this, is that maybe the first file will contain on 
> average the faster units to compile, and the second file contains the 
> longer files to compile. What would happen in this case, is that the first 
> instance finishes first, and so half of a twin processor machine would be 
> unused until the second instance finished. So, my idea was to have the two 
> compiler instances that are "fed" files as fast as they finish compiling 
> them, thereby using full-power until all the files were built.
>
> Jonathan
>