[Libre-soc-bugs] [Bug 132] SIMD-like nmigen signal for partitioning

bugzilla-daemon at libre-soc.org bugzilla-daemon at libre-soc.org
Sat Oct 2 17:10:30 BST 2021


--- Comment #49 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
nggh i just realised, the reversed function __radd__ etc need to be a lot
more sophisticated, because they can take a Signal (or Const, or any AST
Expression) on the rhs, but the return result must be a PartitionedSignal
otherwise what is the thing doing??

which means that where in some cases the underlying code (PartitionedShift)
assumes that it is the *thing being shifted* that may be scalar here
we have a situation where it is the *shift amount* that is scalar.

or maybe that is the other way round.

either way this needs to be looked at and taken into consideration.
bugreport underway....

the other one (already raised) is __getitem__ and Slice, bug #716 which
also need overriding.  maybe. __getitem__ definitely, Part definitely,
Slice perhaps not because it is i believe supposed to be integers
(compile time) only.  dynamic selection uses Array and we are not
tackling that yet.  too complex.

You are receiving this mail because:
You are on the CC list for the bug.

More information about the libre-soc-bugs mailing list