What did you mean anything but fedora core? I think its the best distro out there. I went with core 2 intstead of 3 because 3 is still a bit too new. Since my burner has issues, somone burned it for me.
Fedora Core has its issues, mainly because of the crappy way they compile the code itself.
Mandrake and SuSE will compile the code using 586 switches (say pc-gnu-linux-586) but Fedora compiles their stuff for 386 procs.
The main problem is this; SuSE and Mandrake's code will run on 586 and newer processors. Fedora Core will run on 386 and above, while it seems that Fedora Core has the advantage, it doesn't really, because the compiled code is essentially deprecated 386 code, while it provides maximum flexibility, who uses a 386 nowadays?
Gentoo users brag about being able to compile code for whatever processor they use, since Gentoo's make.conf file will allow the user to use whatever GCC flags you want, for example, mine uses the march=athlon-xp CPU type, so it will generate code that is optimized to my processor. It's a nice feature, but Gentoo's price is a difficult install.
personally, I'd rather stick with Mandrake or SuSE, they have a loyal following, good support, and overall it's one of the more popular distros. While I digress from recommending Fedora, if you're comfortable with it, then by all means go ahead and use it.
A week ago, I would probably have agreed, but not anymore. Here is why....
A bunch of the guys in ACM had the NQueens assignment, so we all wrote up different implementations and tested their speed on 15 queens (which my Java version won :50s on a 800mhz Athlon - 15s on my home machine). A few days later, two of us sat down and wrote both C and CPP versions of my Java version. Very interesting results....
g++ 3.2.? 44s
java 1.4.2 50s
gcc 3.2.? 65s
That's right. The C code was not only slower, but much slower. The guy I was going it with is a big fan of C and was totally blown away (and I've been busting jokes on him for the past week now). Getting back to your point, when he specified -O3 and -march for the P3, it actually ran slower than before (w/ just O3)! Nothing seemed to help. Interestingly enough, compiling the c code with g++ turned in a 44s time too.
Later we installed GCC-3.4.3 and things have gotten a bit stranger. When it decides to run correctly, C is the fastest and puts in a 36s to 38s time. However, it occasionally goes haywire and puts out a 90s plus time (about 1 in 5). Less frequently, it doesn't run correctly.... it finishes w/o crashing but appears to have optimized out the results (probably 1 or 2 in 10). Don't get me started on GCJ either.
So to the whole optimized code thing.... I wouldn't put a whole lot of faith in it. We need to play around more and be more meticulous with our little experiement/study, but there are some funny things going on with GCC at times.