Skip to content

0.1.1 — 16 bit support

Compare
Choose a tag to compare
@proger proger released this 11 Jan 15:24
· 33 commits to main since this release
ad0dbfd

This package adds support for float16 and bfloat16 through templating the warp kernel. Below is the plot for max abs errors comparing the reference implementation and the kernel:

image