Dec 152017
| Computes the overhead that your TSR’s place on your system. | |||
|---|---|---|---|
| File Name | File Size | Zip Size | Zip Type |
| OVERHEAD.C | 1700 | 778 | deflated |
| OVERHEAD.DOC | 2171 | 1050 | deflated |
| OVERHEAD.EXE | 20022 | 13380 | deflated |
Download File OVERHEAD.ZIP Here
Contents of the OVERHEAD.DOC file
OVERHEAD.DOC
------------
OVERHEAD simply wastes a lot of CPU time and executes a lot of interrupts.
It is completely CPU bound, no I/O is done. This is a harmless program,
it's not a destructive diagnostic. I have included the MSC 4.0 source
code so you can see what's going on.
NOTE:This program assumes an IBM-compatable machine as it executes
DOS and BIOS interrupts and looks at the BIOS clock count kept
in low memory.
The purpose is to see how long it takes to perform a given amount of CPU
work. When no TSR's are loaded your program gets as much CPU time as
DOS can manage. But when TSR's are loaded they take some CPU time to
perform their housekeeping. This program lets you see how much time is
being taken by TSR's.
To run the program just type OVERHEAD at the DOS command prompt. The
program will sign on and tell how many iterations it will do. The
defualt is 4 iterations which take about 25 seconds on my AT. To
perform a different number of iterations, add the number you want to the
command line: "OVERHEAD 1" will do just one iteration. When the program
finishes it will show the elapsed time in seconds.
The general approach is to run the program twice. First run it with no
TSR's installed, this is the 'normal' execution time. Then set up your
system as usual with all your favorite TSR's and run the program again.
The difference in run time is caused by the overhead added by the TSR's.
You may be quite surprised by the results.
On my system the 'normal' run time is 25 seconds. With Sidekick and
Superkey installed the run time increased to 34 seconds!! This may make
you wonder if the decrease in overall performance is worth the
convenience of these kinds of programs.
Closing comments:
This isn't meant to be an exhaustive or sophisticated test. It just
does a lot of CPU work and excersises the interrupts that most TSR's like
to take over. It seems to give a general feel for the overhead caused
by TSR's. I've included the source code so you can see exactly what the
program does and you can modify it if you want.
Ed Ross
Compuserve [75776,151]
------------
OVERHEAD simply wastes a lot of CPU time and executes a lot of interrupts.
It is completely CPU bound, no I/O is done. This is a harmless program,
it's not a destructive diagnostic. I have included the MSC 4.0 source
code so you can see what's going on.
NOTE:This program assumes an IBM-compatable machine as it executes
DOS and BIOS interrupts and looks at the BIOS clock count kept
in low memory.
The purpose is to see how long it takes to perform a given amount of CPU
work. When no TSR's are loaded your program gets as much CPU time as
DOS can manage. But when TSR's are loaded they take some CPU time to
perform their housekeeping. This program lets you see how much time is
being taken by TSR's.
To run the program just type OVERHEAD at the DOS command prompt. The
program will sign on and tell how many iterations it will do. The
defualt is 4 iterations which take about 25 seconds on my AT. To
perform a different number of iterations, add the number you want to the
command line: "OVERHEAD 1" will do just one iteration. When the program
finishes it will show the elapsed time in seconds.
The general approach is to run the program twice. First run it with no
TSR's installed, this is the 'normal' execution time. Then set up your
system as usual with all your favorite TSR's and run the program again.
The difference in run time is caused by the overhead added by the TSR's.
You may be quite surprised by the results.
On my system the 'normal' run time is 25 seconds. With Sidekick and
Superkey installed the run time increased to 34 seconds!! This may make
you wonder if the decrease in overall performance is worth the
convenience of these kinds of programs.
Closing comments:
This isn't meant to be an exhaustive or sophisticated test. It just
does a lot of CPU work and excersises the interrupts that most TSR's like
to take over. It seems to give a general feel for the overhead caused
by TSR's. I've included the source code so you can see exactly what the
program does and you can modify it if you want.
Ed Ross
Compuserve [75776,151]
December 15, 2017
Add comments