View Contents of Intel Floating-Point Registers (OS/2, Windows)

When you are debugging at the source code level (by using step commands) you may find that statements involving floating-point operations do not appear to affect the floating-point registers on Intel processors. These registers, STR0 to STR7 in the Registers view, may show invalid or blank contents as you step over floating-point statements.

Because of the Intel floating-point unit architecture and the way source language statements are compiled into machine language, the end of each statement involving floating-point arithmetic involves storing floating-point results to memory. To see the contents of floating-point registers as you step through floating-point instructions, do the following:

  1. Switch your Source view to a Disassembly view, or open an additional pane, and set its view to Disassembly.
  2. Open a registers view, if one is not already open. In the Registers view, expand the Floating-Point Registers section.
  3. Use machine step commands (accessible from the floating dockable toolbar for each debugger process) to step over each disassembly instruction involving floating-point registers. As you step through the disassembly, you will see the contents of floating-point registers change.

You may find the Stack frame and Storage variables views more useful for monitoring floating-point variables, as you can always see what value is assigned to a given variable or at a given address.



Summary of Step Commands
Source-Level Step Commands


Change Representation of Register Contents
Choose Which Registers Display in a Registers View
Step Through Machine Language (Disassembly) Code