Are There Libraries for AVX-512 Opmask Registers?

2 replies · 2K views · last reply
Messages
38,158
Reaction score
10,769
Greg Bernhardt submitted a new blog post

AVX-512 Assembly Programming: Opmask Registers for Conditional Arithmetic
Three horizontal rows of cells under the heading Guide: the top row shows integers 9, -2, 8, 7, -3, 12, 0, 18, -2, -2, 4, 35 with some cells shaded orange; the middle row shows all zeros; the bottom row shows 1s and 0s.


Continue reading the Original Blog Post.
 
Attachments
  • Three horizontal rows of cells under the heading Guide: the top row shows integers 9, -2, 8, 7, -3, 12, 0, 18, -2, -2, 4, 35 with some cells shaded orange; the middle row shows all zeros; the bottom row shows 1s and 0s.
    AVX-512_registers.png
    5.1 KB · Views: 746
  • Like
Likes   Reactions: Greg Bernhardt
I'm learning a lot. Thanks for sharing.

Are there libraries that make use of this instruction set to handle arbitrary length vectors?
 
anorlunda said:
Are there libraries that make use of this instruction set to handle arbitrary length vectors?
Thanks for taking the time to have a look at my article! The conclusion of this article is in the pipeline, and should be published in a few days.

Regarding your question, not that I know of, but then I don't really know what products are using this technology. Most of my efforts have been spent in just figuring out how to use this stuff. The AVX-512 instruction set has about a dozen subsets, some of which are yet to be released. One of the yet-to-be-released subsets is AVX-512VNNI (vector neural network instructions), and is designed to accelerate convolutional neural network algorithms (https://en.wikichip.org/wiki/x86/avx512vnni).

As far as who else is using this technology, the AVX-512 stuff is pretty new, so, and makes some heavy hardware demands. I think the situation is a bit like when Intel and AMD came out with 32-bit processors in the mid-80s, but it took about 10 years for OSes to catch up.
 
Last edited by a moderator: