Follow me on TwitterMy Tweets
Category Archives: Floating Point
A few months ago I saw a blog post touting fancy new SSE3 functions for implementing vector floor, ceil, and round functions. There was the inevitable proud proclaiming of impressive performance and correctness. However the ceil function gave the wrong … Continue reading
Is IEEE floating-point math deterministic? Will you always get the same results from the same inputs? The answer is an unequivocal “yes”. Unfortunately the answer is also an unequivocal “no”. I’m afraid you will need to clarify your question. My … Continue reading
Last year I pointed out that float variables can be converted to text and then back to the same binary value using printf(“%1.8e”). I also supplied a test program that used C++ 11 threading to quickly prove this claim on … Continue reading
The end of the year is a traditional time for lists and I thought I’d do a post just to summarize the most interesting or surprising things that I discovered while writing this blog during 2012.
This is for references, code examples, and discussion regarding the floating-point article in the October 2012 Game Developer Magazine.
I’ve seen a few online discussions linking to my Comparing Floating Point Numbers page for misguided reasons and I wanted to discuss those reasons to help people understand why throwing epsilons at the problem without understanding the situation is a … Continue reading
I gave a talk at iFest on Saturday that briefly covered some performance and precision issues taken from my series of floating-point blog posts. My talk was after Ed Fries, who did a brilliant talk on art and constraints, centered … Continue reading
Denormals, NaNs, and infinities round out the set of standard floating-point values, and these important values can sometimes cause performance problems. The good news is, it’s getting better, and there are diagnostics you can use to watch for problems. In … Continue reading
Floating-point math has an answer for everything, but sometimes that’s not what you want. Sometimes instead of getting an answer to the question sqrt(-1.0) (it’s NaN) it’s better to know that your software is asking imaginary questions. The IEEE standard … Continue reading