home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
OS/2 Shareware BBS: 8 Other
/
08-Other.zip
/
OS2TEST.ZIP
/
OS2TEST.DOC
next >
Wrap
Text File
|
1988-05-18
|
2KB
|
38 lines
Real/Protected/Multitask Comparisons
(For OS/2 owners who can't do much with it yet)
J.P.Davis
5/11/88
Having recently started to play with IBM OS/2 v1.0 and Microsoft Basic
Compiler 6.0, which permits compiling to "real" or "protected" mode, I
thought I'd try some speed tests. A simple formula is run in a continuous
loop and number of iterations, time, and iterations per unit time are
printed to screen at the user's prompt. Thus, real vs protected vs
both vs multi-copies may be looked at.
Note that the protected.exe will only load in the OS/2 mode and the real.exe
only in DOS mode. Further, multi-copies of protected.exe may be loaded and
run concurrently but only one copy of real.exe may be loaded since OS/2 has
but one "DOS BOX". However, the real.exe may be moved from foreground
to background and reverse. Remember that when the real.exe is moved to
background its running is SUSPENDED, in contrast to protected.exe which
keeps running in either foreground OR background - but at different speeds.
A warning. Don't place more significance on speed results for these various
combinations than you should. For example, string manipulations may
respond differently than the little number-cruncher used in these programs.
Also, let the individual runs go for at least 10 seconds or so. Beyond
that time increment, the iterations per unit time should have settled down
to a reasonably constant value.
R-TEST.BAS and P-TEST.BAS are identical except for the titles. The .exe's
were compiled from the command line using either real or protected mode and
used the /EX(epack) option with the Microsoft Basic Compiler 6.0.
The above programs use a DO:LOOP WHILE INKEY$="". OS2 seems to be exception-
ally sluggish compared to DOS in looping through this keyboard input test.
If a FOR-NEXT LOOP is used OS2 is on a speed par with DOS. Thus, general-
izations on speed are VERY suspect for OS2/DOS type comparisons and depend
closely on the specifics of the test.