rms@AI.MIT.EDU (05/12/89)
How many people would like to see built-in support for transcendental functions in GCC? This would mean adding new rtl codes, new optabs, new standard pattern names, and adding numerous patterns to all the machine descriptions to take advantage of the feature. I hesitate to install so much change unless there is real enthusiasm for it. Note that the built-in functions would fail to set errno and therefore would not be compatible with ANSI C.
rds95@leah.Albany.Edu (Robert Seals) (05/12/89)
In article <8905112047.AA00256@sugar-bombs.ai.mit.edu>, rms@AI.MIT.EDU writes: > How many people would like to see built-in support for transcendental > functions in GCC? At least 1 - me. (I wasn't sure whether to post a followup or reply direct - hope I was right) rob
kaul@icarus.eng.ohio-state.edu (Rich Kaul) (05/12/89)
In article <8905112047.AA00256@sugar-bombs.ai.mit.edu> rms@AI.MIT.EDU writes: >How many people would like to see built-in support for transcendental >functions in GCC? ... >Note that the built-in functions would fail to set errno and therefore >would not be compatible with ANSI C. I would like to see it, but only if it were an option. The increase in speed would be helpful to those of us who tend to use those functions quite a bit. -rich -=- Rich Kaul | "Of course, America had often been The Ohio State University | discovered before Columbus, but it 2015 Neil Ave, Columbus, OH 43210 | had always been hushed up." kaul@icarus.eng.ohio-state.edu | -Oscar Wilde