BUG: Using Intrinsic Versions of log10 and fmod (152056)

The information in this article applies to:

Microsoft Visual C++ 4.0
Microsoft Visual C++ 4.1
Microsoft Visual C++ 4.2
Microsoft Visual C++, 32-bit Enterprise Edition 5.0
Microsoft Visual C++, 32-bit Professional Edition 5.0

This article was previously published under Q152056

SYMPTOMS

In some cases, using the intrinsic versions of log10 and fmod together in a program may produce incorrect results. The sample code below illustrates the problem.

When the result of log10 is a negative whole number that is divided by 1.0 using fmod, the result of fmod will be -1.0 rather than 0.0. This only occurs when the compiler generates the intrinsic versions of these functions. You need to use the /Oi compiler option to have the compiler generate intrinsic versions of functions.

RESOLUTION

When using log10 and fmod together in a function, as in the sample code, do not use the /Oi compiler option.

STATUS

Microsoft has confirmed this to be a bug in the Microsoft products listed at the beginning of this article. We are researching this problem and will post new information here in the Microsoft Knowledge Base as it becomes available.

MORE INFORMATION

Sample Code

   /* Compile options needed: /Oi
   */ 
   #include <math.h>
   #include <stdio.h>

   void main()
   {
       double dValue, theLog, theMod;

       dValue = 0.01;
       theLog = log10( dValue );
       theMod = fmod( theLog, 1.0 );

       printf( "The log10 of %f is %f\n", dValue, theLog );
       printf( "The fmod of (%f,1.0) is %f\n", theLog, theMod );
       printf( "for  all x, fmod(x, 1.0) should always be 0.0\n" );
   }

Modification Type:	Major	Last Reviewed:	12/10/2003
Keywords:	kbBug kbCodeGen kbCompiler KB152056