News for hackers. Stuff that matters.

A comment on Float Toy. Parent comment

Re:Float Toy by Sharlin on Saturday August 20, 2022 @08:34PM (#32536347)
I definitely don't think half precision is completely normal obvious type except maybe in certain circles. None of current mainstream languages support it as a primitive type, for one. No mainstream CPUs have hardware support for half precision.
[ Parent | on HN ]
Re:Float Toy by ChrisLomont on Saturday August 20, 2022 @10:11PM (#32536810)
C# supports it. CUDA supports it. ARM and related C/C++ compilers support it. Intel is adding hardware support in upcoming chips, so expect a lot more languages to add it.
[ Parent | on HN ]
Re:Float Toy by adgjlsfhk1 on Sunday August 21, 2022 @02:15AM (#32537831)
Julia also supports it. it led to a really funny graph when doing performance tests on Fujitsu, because Julia is the only language that supports fp16 and is fast enough to write blas in, so there were some graphs where it was the only entry because c and fortran didn't bother showing up.
[ Parent | on HN ]
Re:Float Toy by TazeTSchnitzel on Sunday August 21, 2022 @05:24AM (#32538583)
Most Intel and AMD CPUs produced in the last ~10 years have hardware support for converting 16-bit floats to and from 32-bit. That's not full hardware support of course, but you don't need it to be able to make good use of them. https://en.wikipedia.org/wiki/F16C
[ Parent | on HN ]
Re:Float Toy by Sharlin on Sunday August 21, 2022 @01:38PM (#32542458)
Except when it comes to SIMD. Having twice the lanes compared to f32 would be a real benefit.
[ Parent | on HN ]