News for hackers. Stuff that matters.

A comment on FloatX: A C++ Library for Customized Floating-Point Arithmetic.

Re:FloatX: A C++ Library for Customized Floating-Point Arith... by bhuthesh_r on Wednesday January 08, 2020 @06:51AM (#21990555)
I'm trying to use 16 bit floats for matrix multiplication on x86-64. I found solutions for ARM and some NVIDIA GPUs but none for any X86-64 chips. Any pointers in this direction would be helpful.
[ Parent | on HN ]
Re:FloatX: A C++ Library for Customized Floating-Point Arith... by integricho on Wednesday January 08, 2020 @07:25AM (#21990754)
Here is a little sample code I threw together, it shows the whole cycle, from conversion to half-floats to the conversion back to floats and performing a simple multiplication of the values:

https://godbolt.org/z/FYu_rK

Hope it helps.

[ Parent | on HN ]
Re:FloatX: A C++ Library for Customized Floating-Point Arith... by bhuthesh_r on Thursday January 09, 2020 @02:09AM (#21999137)
Thank you.
[ Parent | on HN ]
Re:FloatX: A C++ Library for Customized Floating-Point Arith... by helltone on Wednesday January 08, 2020 @04:15PM (#21995877)
https://github.com/google/gemmlowp
[ Parent | on HN ]