In the context of bitwise replicability, which always refers to floating-point computations, the main constraint is that floating-point arithmetic is incompletely specified in most of today’s programming languages, and that whatever specification there is is incompletely implemented in many of today’s compilers.
I’m confused by this, IEEE754 has has been around for a long time. As long as you’re using the same order of operations to allow for non-associativity and same size floats for the same rounding you should get the same result.
Since every compiler optimizes differently, the same program source code yields different results on different platforms.
I’d argue that any compiler that does this out of the box to floating point numbers is buggy, as it has changed behaviour.
I’m confused by this, IEEE754 has has been around for a long time. As long as you’re using the same order of operations to allow for non-associativity and same size floats for the same rounding you should get the same result.
I’d argue that any compiler that does this out of the box to floating point numbers is buggy, as it has changed behaviour.
It’s not quite so simple:
http://gafferongames.com/networking-for-game-programmers/floating-point-determinism/ https://randomascii.wordpress.com/2013/07/16/floating-point-determinism/