kernel/ksupport: add linalg functions#2527
Conversation
| api!(yn), | ||
|
|
||
| // linalg | ||
| api!(np_linalg_matmul = ::linalg::np_linalg_matmul), |
There was a problem hiding this comment.
Again I don't think a firmware call and Rust crate are necessary for a couple for loops with a few additions and multiplications in them.
There was a problem hiding this comment.
Also this is already implemented via the @ operator so AFAICT you can just remove this entirely.
There was a problem hiding this comment.
Also this is already implemented via the
@operator so AFAICT you can just remove this entirely.
Should I also remove np_linalg_det as that too can be implemented in a straight-forward manner.
There was a problem hiding this comment.
I think determinant isn't that straightforward and there are a bunch of different algorithm? Maybe using the same crate that does the other operations is fine here.
AR-PyT
force-pushed
the
add_linalg_functions
branch
from
0795195 to
4e37ac4
Compare
|
Is this tested? I don't think this is going to work without adding nalgebra to cargo. |
2 participants
ARTIQ Pull Request
Description of Changes
Adds runtime implementation of
np.linalg/sp.linalgfunctions (M-Labs/nac3#478 and M-Labs/artiq-zynq#309)Depends on #2526
Related Issue
Type of Changes
Steps (Choose relevant, delete irrelevant before submitting)
All Pull Requests
Code Changes
flake8to check code style (follow PEP-8 style).flake8has issues with parsing Migen/gateware code, ignore as necessary.Licensing
See copyright & licensing for more info.
ARTIQ files that do not contain a license header are copyrighted by M-Labs Limited and are licensed under LGPLv3+.