home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Kit PC World De Ampliacion De Windows 95
/
Kit PC World de ampliacion de Windows 95.iso
/
clarion
/
gauge
/
gauge.z
/
GAUGE.TXT
< prev
next >
Wrap
Text File
|
1995-10-13
|
3KB
|
67 lines
The Clarion for Windows Performance Gauge
GAUGE launches test programs that compare the relative performance
of Clarion for Windows, Delphi, Visual Basic, and PowerBuilder.
Launching the Benchmarks
The benchmark programs are stand-alone .EXEs that can be run
independently. However, these programs have been written to
operate as dynamic data exchange (DDE) servers. The Performance
Gauge is their client, launching them with a RUN statement, and
stimulating them to action using the Clarion DDE library.
The Clarion and Delphi benchmark programs are complete "one-piece"
executables. The Visual Basic and PowerBuilder benchmarks require
run-time .DLL files. The resulting executable sizes can be
displayed by pressing the ".EXE Size..." button.
Though not measured by the Performance Gauge, the time required to
load a VB program load time is considerably greater than that
required for Clarion or Delphi programs. Because of their huge
size, PowerBuilder programs take a very long time to load.
The Sieve of Eratosthenes Benchmark
The "Sieve of Eratosthenes" is an ancient algorithm for finding
the number prime numbers between one and any positive integer.
This algorithm is commonly used to benchmark compiler efficiency.
The Performance Gauge measures the number of times per minute that
each program can count the primes between 1 and 1,000. To produce
similar duration, the Clarion "Sieve" cycles 4,000 times, Delphi
cycles 2,000 times, VB cycles 100 times, and PB cycles 7 times.
The Video Performance Benchmark
To test video performance, a typical 3 dimensional dialog box is
displayed repeatedly. The dialog box contains text controls, a
combo box, a spin box, radio buttons, check boxes, and "OK" and
"Cancel" buttons. All entry fields are initialized with data.
As soon as the dialog box has been completely painted, it is
closed. When the window is completely removed, it is redisplayed.
A window is removed by repainting the window beneath it. In order
to control this process, the benchmark dialog is displayed over a
maximized window.
The "UpdateWindow" function from the standard Windows API is used
to verify that the complete Window is visible and that it has been
completely removed. In Clarion this is unnecessary (but benign)
because the ACCEPT loop is not initiated until the window has been
painted. In Delphi, the "Update" method performs an "UpdateWindow".
Visual Basic requires custom controls (.VBXs) to produce 3D
effects. 3D Gizmos by MicroHelp have been used for this purpose.
To produce similar duration, the Clarion benchmark cycles 25
times, Delphi cycles 40 times, VB cycles 20 times, and PB cycles
10 times.
Installation
The Performance Gauge can be run from any drive and path, however,
the directory containing CWGAUGE.EXE must also contain Delphi, VB,
and PB sub-directories for the benchmark executables.