SAMPLE: Floating-Point Samples for Windows (97265)

MORE INFORMATION

The following file is available for download from the Microsoft Download Center:
For additional information about how to download Microsoft Support files, click the following article number to view the article in the Microsoft Knowledge Base: Microsoft scanned this file for viruses. Microsoft used the most current virus-detection software that was available on the date that the file was posted. The file is stored on security-enhanced servers that help to prevent any unauthorized changes to the file.

• FLOATAPP.EXE: Windows-based application used as a shell for testing application and dynamic-link library (DLL) floating-point issues. This application tests floating-point speed, generates exceptions, installs signal handler, masks and unmasks exceptions, and communicates with WIN87EM.DLL. Some capabilities depend on the math package chosen and/or the presence of a math coprocessor.
• FLOATDLL.DLL: Floating-point Windows DLL. FLOATDLL is a subset of FLOATAPP, to which it dynamically links. FLOATDLL can generate and handle floating-point exceptions, and can be configured to install and maintain its own signal handler or use the default (or application) handler.
• FLOATDOS.EXE: Floating-point MS-DOS(R) application. FLOATDOS is a subset of FLOATAPP that runs outside of Windows (or in a "DOS box"), thus allowing you to compare an MS-DOS application with a Windows-based application for exception handling and (most importantly) calculation speed.
• HIDE87.COM: MS-DOS terminate-and-stay-resident (TSR) program that hides the 80x87 from Windows-based applications. HIDE87 must be installed before you start Windows so you can trick Windows into thinking there's no math coprocessor.
• SHOW87.EXE: MS-DOS control program that disables the HIDE87 TSR. SHOW87 provides a way to turn off HIDE87 without rebooting the machine. As with HIDE87, SHOW87 must be executed before you start Windows so it can detect the math coprocessor.

See the sample code and the "Floating Point in Windows" technical article for descriptions of these programs, their coding techniques, and the Windows floating point architecture. For information on building these programs and discussions of C version 6.0 versus C/C++ version 7.0 and emulator versus alternate math packages, see the make file. Note that these sample have major limitations if compiled with C 6.0.

WINFLOAT was built and tested under Microsoft Windows version 3.1 using Microsoft Macro Assembler version 6.0 and Microsoft C/C++ version 7.0.