Compile CUDA code on 64 bit machine

I 'm trying to compile some CUDA code on a 64 bit computer, and the error message is
/usr/bin/ld: skipping incompatible /home/me/cuda/lib/ when searching for -lcuda

Is there a solution to this problem? Thanks!

CUDA is not currently supported in 64bit operating systems. The next release will provide 64bit Linux support.

Ok, thank you very much.

You can use -m32 (or -Xcompiler -m32) during compilation and linking (-Xlinker -m32 if you don’t use gcc directly but through nvcc). It will run, but the emulation code gives very strange outputs on my Fedora 6 on an x86_64.
Somebody any ideas about running the emulator on an x86_64?
Hopefully the 64bit version will be released soon ;-)

Just quickly - I had no trouble compiling & linking on FC6 x86_64, the -m32 flag was already in the 0.8 release, had to show it where one of the 32 bit libs was and that was all (see Linux forum for lots of examples). Emulation works but it is very slow - I get 4% user time, 30% system time on a dual core. Problem in pthreads? Could also be a gcc 4.1 optimiser problem. Not supported so no complaints. Waiting for next release.

Oops, the strange output came from a silly mistake. When will I learn not to do difficult tasks in the middle of the night ;-)