4 ms·

I've never heard anyone say any of these supposed myths, except for the first one, sort of, but nobody means what the first one pretends it means, so this whole

I've never heard anyone say any of these supposed myths, except for the first one, sort of, but nobody means what the first one pretends it means, so this whole post feels like a big strawman to me.

rollcat 1y ago
People do say a lot of nonsense about floats, my younger self trying to bash JavaScript included. They e.g. do fine as integers - up to 2**53; this could be optimised by a JIT to use actual integer math.
Cieric 1y ago
Just to add some context Adam works at AMD as a dev tech, so he is constantly working with game studios developers directly. While I can't say I've heard the same things since I'm somewhere else, I have seen some of the assumptions made in some shader code and they do line up with the kind of things he's saying.
jbjbjbjb 1y ago
I was putting together some technical interview questions for our candidates and I wanted to see what ChatGPT would put for an answer. It told me floating point numbers were non-deterministic and wouldn’t back down from that answer so it’s getting it from somewhere.
AshamedCaptain 1y ago
I concur, it seems low effort, and the only real common "myth" (the 1st one) is not really disproven. Infact the very example he puts goes to prove it, as it is going to become an infinite loop given large enough N.... Also compiler optimizations should not affect the result. Specially without "fast-math/O3" (which arguably many people stupidly use nowadays, then complain).
- zokier 1y ago
  > Also compiler optimizations should not affect the result. Specially without "fast-math/O3" (which arguably many people stupidly use nowadays, then complain). -ffp-contract is annoying exception to that principle.
jandrese 1y ago
The first rule is true, but relying on it is dangerous unless you are well versed in what floats can and can not be represented exactly. It's best to pretend it isn't true in most cases to avoid footguns.
janalsncm 1y ago
Same, but I still learned a fair amount anyways. I say no harm, no foul.
nspattak 1y ago
i am laughing and crying at the same time... a couple of decades ago, I was still a lowly applied mathematics (and aspiring programmer) student when i proposed to a professor to change the compiler options from O2 to O3 to gain performance (reducing run time by a significant percentage like 50% or more...) only to be disregarded by saying "no, compiler optimizations break numerical accuracy".

I've never heard anyone say any of these supposed myths, except for the first one, sort of, but nobody means what the first one pretends it means, so this whole post feels like a big strawman to me.

rollcat 1y ago

People do say a lot of nonsense about floats, my younger self trying to bash JavaScript included. They e.g. do fine as integers - up to 2**53; this could be optimised by a JIT to use actual integer math.

Cieric 1y ago

Just to add some context Adam works at AMD as a dev tech, so he is constantly working with game studios developers directly. While I can't say I've heard the same things since I'm somewhere else, I have seen some of the assumptions made in some shader code and they do line up with the kind of things he's saying.

jbjbjbjb 1y ago

I was putting together some technical interview questions for our candidates and I wanted to see what ChatGPT would put for an answer. It told me floating point numbers were non-deterministic and wouldn’t back down from that answer so it’s getting it from somewhere.

AshamedCaptain 1y ago

I concur, it seems low effort, and the only real common "myth" (the 1st one) is not really disproven. Infact the very example he puts goes to prove it, as it is going to become an infinite loop given large enough N.... Also compiler optimizations should not affect the result. Specially without "fast-math/O3" (which arguably many people stupidly use nowadays, then complain).

zokier 1y ago
> Also compiler optimizations should not affect the result. Specially without "fast-math/O3" (which arguably many people stupidly use nowadays, then complain). -ffp-contract is annoying exception to that principle.

jandrese 1y ago

The first rule is true, but relying on it is dangerous unless you are well versed in what floats can and can not be represented exactly. It's best to pretend it isn't true in most cases to avoid footguns.

janalsncm 1y ago

Same, but I still learned a fair amount anyways. I say no harm, no foul.

nspattak 1y ago

i am laughing and crying at the same time... a couple of decades ago, I was still a lowly applied mathematics (and aspiring programmer) student when i proposed to a professor to change the compiler options from O2 to O3 to gain performance (reducing run time by a significant percentage like 50% or more...) only to be disregarded by saying "no, compiler optimizations break numerical accuracy".