Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Someone’s Been Messing with My Subnormals

moyix.blogspot.com

Post reply on HN
Print view View original

31–40 of 132 posts

Re: Someone’s Been Messing with My Subnormals

#31

Post by mananaysiempre »

Following the article’s links, I fail to find an actual example of anything failing to converge in flush-subnormals mode. I mean, I’m sure one could be squeezed out, but the justification given amounts to “Sterbenz’s lemma [the one that rephrases “catastrophic cancellation” as “exact differences”] fails, maybe something somewhere also will”. And my (shallow but not nonexistent) experience with numerical analysis is that proofs lump subnormals with underflow, and most of them don’t survive even intermediate underflows.

(AFAIU the original Intel justification for pushing subnormals into 754 was gradual underflow, i.e. to give people at least something to look at for debugging when they’ve ran out of precision.)

So, yes, it’s not exactly polite to fiddle with floating-point flag bits that are not yours, and it’s better that this not happen for reproducibility if nothing else, but I doubt it actually breaks any interesting numerics.

Re: Someone’s Been Messing with My Subnormals

#32

Post by cesarb »

At a previous company I worked at, we had an issue with our software (Windows-based, written in a proprietary language) randomly crashing. After some debugging, we found that this happened whenever the user made some specific actions, but only if, in that session, the user had previously printed something or opened a file picker. The culprit was either a printer driver or a shell extension which, when loaded, changed the floating point control word to trap. That happened whenever the culprit DLL had been compiled by a specific compiler, which had the offending code in the startup routine it linked into every DLL it produced.

Our solution was the inverse of the one presented in this article: instead of wrapping our routines to temporarily set the floating point control word to sane values, we wrapped the calls to either printing or the file picker, and reset the floating point control word to its previous (and sane) value after these calls.

Re: Someone’s Been Messing with My Subnormals

#34

Post by klysm »

cesarb wrote: post #32

At a previous company I worked at, we had an issue with our software (Windows-based, written in a proprietary language) randomly crashing. After some debugging, we found that this happened whenever the user made some specific actions, but only if, in that session, the user had previously printed something or opened a file picker. The culprit was either a printer driver or a shell extension which, when loaded, changed…

That is one hell of a war story - I didn't realize that kind of failure was even possible, but it is truly terrifying.

Re: Someone’s Been Messing with My Subnormals

#35

Post by trelane »

actually_a_dog wrote: post #25
nsajko wrote: post #21

-Ofast isn't a good name for the option, but in GCC's defense the manual is pretty clear about all this, and there's no excuse for blindly turning on compiler options - they literally change the semantics of your code.

I agree. I think --ffast-math should actually be called --finexact-math. One would also hope that explicitly disabling an option on the command line would, you know, explicitly disable the option, but maybe that's too much to ask.

-fyolo-math?

-fgoodenough-math?

-fbroken-but-fast-math

Re: Someone’s Been Messing with My Subnormals

#36

Post by moyix »

mananaysiempre wrote: post #31

Following the article’s links, I fail to find an actual example of anything failing to converge in flush-subnormals mode. I mean, I’m sure one could be squeezed out, but the justification given amounts to “Sterbenz’s lemma [the one that rephrases “catastrophic cancellation” as “exact differences”] fails, maybe something somewhere also will”. And my (shallow but not nonexistent) experience with numerical analysis is t…

The gevent issue has an example:

https://github.com/gevent/gevent/pull/1820

I haven't examined the code of scipy.stats.skellam.sf so I can't say for sure that it's not converging, but it's clearly some kind of pathological behavior.

Re: Someone’s Been Messing with My Subnormals

#39

Post by jcranmer »

bee_rider wrote: post #15

A decorator is a nice idea for this. I was going to suggest another package that just resets the MXCSR when imported, but I guess... hypothetically... some function might actually want the FTZ behavior.

If you want that behavior, you should explicitly enable it/disable it at the borders of the region where you want that behavior, rather than screwing over everybody for your own benefit.

Re: Someone’s Been Messing with My Subnormals

#40

Post by pavlov »

klysm wrote: post #34
cesarb wrote: post #32

At a previous company I worked at, we had an issue with our software (Windows-based, written in a proprietary language) randomly crashing. After some debugging, we found that this happened whenever the user made some specific actions, but only if, in that session, the user had previously printed something or opened a file picker. The culprit was either a printer driver or a shell extension which, when loaded, changed…

That is one hell of a war story - I didn't realize that kind of failure was even possible, but it is truly terrifying.

Direct3D used to flip the x87 FPU to single precision mode by default. This produced some amazing bugs when your other C libraries reasonably assumed that a double would be at least 64 bits. (The FPU mode settings affected the thread that called Direct3D, and most programs used to be single-threaded.)

It seems they changed this behavior in Direct3D 10:

https://microsoft.public.win32.programmer.directx.graphics.n...

Post reply on HN