Sets the tolerance used by default when checking the equality of floating point values within the test assembly, fixture or method.