Is bfloat16 ever used for graphics?

[tobimarsh]

Bfloat 16 is a half precision floating point format that has the same 8-bit exponent as single precision, but only 7 (plus 1 implied) bits of significant. Surprisingly, this turns out to be adequate precision for many machine learning applications, so a lot of resources are being put into making arithmetic in this format run fast.

 

Given that, it would seem to make sense to also try to use it for graphics. Using it for RGB components during calculation, for example, would allow a much wider dynamic range of light sources to be rendered, compared to just trying to calculate with 8-bit integers. At the same time, it could potentially be faster than using single precision floating point for RGB components.

 

Are any existing graphics rendering systems actually using it for such purposes?

[igormp]

No, there's not. The precision the bfloat16 offers is not enough for that scenario, and the extra dynamic range that it offers isn't needed, that why fp16 is a better pick.

 

But you also do need to keep in mind that most regular GPUs being used nowadays (i.e.: pascal series) have a capped fp16 throughput, so you need all your maths done in fp32 for max performance in consumer products.