🚀 go-pugleaf

olsen@delta.eecs.nwu.edu (Jeff S. Olsen) writes:

>In article <1771@pdxgate.UUCP> dmatlock@eecs.cs.pdx.edu (Delbert Matlock) writes:
>>I'm running Borland C++ on a 386 with 4MB of memory.  I originally set up the
>>PIF file to only give BC 512K of expanded memory.  BC took nearly 20 minutes
>>to compile and link one of the sample windows programs.  When I ran the EMS up
>>to 1.5MB, compile and link time came down to under two minutes.  So, I guess 
>>you could run BC on a 1 Meg machine, but pack a lunch.

>Did you recompile the same program the second time?  If so, did you
>remove any pre-compiled headers that might have been generated during
>the first compile?

All OBJ, EXE, and RES files were removed between the two compilations.  I 
should add, however, that when the EMS was set at 512K the disk worked 
extremely hard.  With the EMS at 1.5MB, disk access was slowed down to an
occasional flash.  It looks like BC will keep almost all of "Cx.LIB" in 
EMS when it is working if it can.

=============================================================================
Delbert Matlock                         Internet:  dmatlock@eecs.cs.pdx.edu
MicroNet Northwest                      
Voice:  (503)228-3071