Inline-math and GCC 2.7.2

Inline-math and GCC 2.7.2

Post by James B. Hill » Fri, 15 Mar 1996 04:00:00


I've got GCC 2.7.2 with  Saw the inline-math package on
Sunsite and grabbed it.

I don't intentionally do a bunch of number crunching with trig functions
and the like, but surely many X apps end up making use of the functions
improved in the package.

Since the package was release in Jul 95, is it still relevant to incorporate
it with GCC 2.7.2 and that particular libm?  Or could it be that the
latest libm might include these fixes?  In short, would the package be
of use, overkill, or dangerous?



1. Inline math function expansion by GCC's optimization?


I have a program which works well under gcc 2.91.66. Today I switched to gcc
2.96, but it says it cannot found pow() function in linking process.

So I tried to play with compiler options to see what happened. My program
was complied with -O2 optimization, and with 2.91.66, the pow() function is
optimized with internal inline functions. In other words, the pow() function
is replaced by internal inline expansion rather than linked against the
standard library.

However, with 2.96, no matter what optimzation level I specify, the pow()
function is never inlined by the compiler.

My question is, is there any compiler option which can turn on math function
inline expansion for 2.96?


2. FS: Digital AlphaStation 600 5/266

3. math inline bugs with egcs1.1.2 glibc2.1.1

4. HTB in relation to TBF

5. (fwd) REPOST: inline-math-1.3

6. Packet dropping

7. REPOST: inline-math-1.3

8. problems

9. "extern inline" to "static inline" allows compile

10. static inline, and extern inline in the kernel source

11. why is "static inline" preferred over "extern inline"?

12. GCC's inline assembler, and it's apparent lack of knowledge

13. Inline assembler with gcc