[comp.windows.ms.programmer] BCX vsBC

jdb@reef.cis.ufl.edu (Brian K. W. Hook) (03/10/91)

Has anyone felt a noticeable difference between running Borland C++ in
and out of protected mode.  I ran some bench marks of the BCC and BCCX
and got times within 10% of each other most of the time:

8K file
included Windows.h
prcompiled headers on

24 SECONDS using BCC 
" "              BCX
13 SECONDS using BCC -Qe or -Qx

This is with 33mhz 386, 8MB of RAM, configured as follows:

2MB disk cache (NCACHE-F EXT=2048)
1MB for DOS (+384K of shadowing)
5MB for whatever else wants it, configured as XMS

I have tried using -Qx, -Qe, BC, BCCX, BCX, BCC, and disabling/enabling a
greater cache and I have YET to notice major increases in speed!

One other note....one of the .PRJ files for Windows that came with it
compiled in 20 minuites!! It ran out of memory in about 1 minute, (before
it finished the include files) and spent the rest of the time swapping.  I
don't even remember which one it was!  This is NTO a good a sign...\

Brian

LIBCRN@BYUVM.BITNET (03/12/91)

=========================================================================

>Has anyone felt a noticeable difference between running Borland C++ in
>and out of protected mode.  I ran some bench marks of the BCC and BCCX
>and got times within 10% of each other most of the time:

>8K file
>included Windows.h
>prcompiled headers on

>24 SECONDS using BCC
>" "              BCX
>13 SECONDS using BCC -Qe or -Qx

>This is with 33mhz 386, 8MB of RAM, configured as follows:

>2MB disk cache (NCACHE-F EXT=2048)
>1MB for DOS (+384K of shadowing)
>5MB for whatever else wants it, configured as XMS

>I have tried using -Qx, -Qe, BC, BCCX, BCX, BCC, and disabling/enabling a
>greater cache and I have YET to notice major increases in speed!

   In the post that I made which asked about running BCX within Windows
(again I apologize for wasting bandwidth on a question that I should have
solved myself by reading the manual), I listed compile-and-link times for
two simple Windows programs.  The difference that I saw was VERY NOTICEABLE.
In case you missed my earlier post, here are the numbers again:


PARTY.C    (from Petzold's "Programming Windows" book):
        BC  --  5.5 minutes.
        BCX --  16 seconds.
WHELLO.CPP (from Borland's sample programs)
        BC  --  8 minutes   (that's right EIGHT MINUTES!)
        BCX --  40 seconds.


   Of course my machine isn't nearly as monstrous as yours so you may not
get the same type of results as I did.  My hardware is much more modest:
        * * IBM PS/2 Model 55sx   with  4 Mbytes of memory. * *

>Brian

--Cory (libcrn@byuvm.bitnet)