[Libre-soc-bugs] [Bug 713] PartitionedSignal enhancement to add partition-context-aware lengths

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=713

--- Comment #51 from Jacob Lifshay <programmerjake at gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #48)
> (In reply to Jacob Lifshay from comment #43)
> > iirc nothing actually uses PartitionedSignal yet, so a API break to using
> > SimdLayout shouldn't be as big a problem.
> 
> except that there may be good reasons why other people (other customers
> whom i cannot reveal or discuss because it is commercially confidential)

can you tell me privately? i was not aware of anyone using it...

I'm kinda inclined to say "you decided to use unstable not even released
software, you reap the consequences of said software changing", but I recognize
that that's probably not helpful...

> who might wish to use PartitionedSignal *as it is* i.e. who may not
> wish to use Elwidths or who may wish to define their own
> adapter or class on *top* of PartitionedSignal, or whom may
> wish to have *all* the mask points available not the subset as
> fictated by Elwidths.
> 
> basically i am saying it is too early to go destroying the
> existing PartitionedSignal API and wholesale converting all
> of the unit tests to ElwidthPartitionedSignal.

Well, if we do work in a separate branch, we can see if the design will pan
out. if not, we can use pieces of the code in PartitionedSignal 3.0 (since
simd_signal is kinda PartitionedSignal 2.0.0.alpha).

I think copying the existing PartitionedSignal and naming the copy
ElwidthSimdSignal is better than having it be a subclass...subclassing will
make a bunch of messy unnecessary interactions. Also, naming it Simd is waay
shorter than Partitioned.

I did kinda run out of time today with getting @deduped to work, so will put
off writing SimdLayout till next week. Since I'm the one pushing for it, I'll
handle the work needed to integrate it.

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