C code for floating point number addition with 100% accuracy.
The test.c program can run about 501 million combinations of floating point numbers, testing every possible exponent with randomly generated significands. In a test of 501,350,400 combinations, the FloatAdd.c program got 0 wrong and proved it's 100% accuracy.