[HN Gopher] OpenGL Machine Learning on Low-End Hardware
___________________________________________________________________
OpenGL Machine Learning on Low-End Hardware
Author : rcarmo
Score : 87 points
Date : 2021-11-14 09:58 UTC (2 days ago)
(HTM) web link (hackaday.com)
(TXT) w3m dump (hackaday.com)
| RattleyCooper wrote:
| They author of the library is trying to sell commercial licenses.
| People should get together, recreate this and slap an MIT license
| on it.
| pjmlp wrote:
| How dare the author to earn money for their efforts, tstst.
| RattleyCooper wrote:
| Oh, I don't care if they do and never said they shouldn't, I
| just think someone should make a FOS version.
|
| How dare someone want to use something free and open source,
| that's developed by a vast community of individuals instead
| of relying on 1 individual (who might disappear at a moment's
| notice) for every single update and bug fix. Tstst.
|
| See, I can be snarky too
| pjmlp wrote:
| I see, so what?
| varelse wrote:
| Almost every programming model is Turing complete therefore you
| can do just about anything with them. But unless OpenGL shader
| language has equivalent concepts of the warp and the syncthreads
| instruction, you end up having to do a global sync every time you
| want to do any synchronization and that means returning to the
| CPU and that's really slow in comparison and therefore very power
| inefficient as well.
___________________________________________________________________
(page generated 2021-11-16 23:01 UTC)