Discussions of the problem...

1) David Goldberg: What Every Computer Scientist Should Know About Floating-Point Arithmetic. ACM Computing Surveys, Vol. 23, No. 1 (March 1991), pages 5–48.
2) GSL IEEE Floating Point Arithmatic
3) Floating Point Determinism
4) A question/answer on stackoverflow...
5) Random Ascii Blogs
6) Consistency: How to Defeat the Purpose of IEEE Floating Point

Some possible pieces to a solution

1) One implementation of a deterministic FP library... (really just a header file which tries to maniupulate IEEE FP options consistently accross machines/compilers).
2) MPFR -- Multiple Precision Floating point library
3) Cross platform math library...