[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #65046] half type (16-bit floating point) supp
From: |
Rik |
Subject: |
[Octave-bug-tracker] [bug #65046] half type (16-bit floating point) support |
Date: |
Tue, 19 Dec 2023 13:00:29 -0500 (EST) |
Update of bug#65046 (group octave):
Priority: 3 - Low => 2
Summary: half type support => half type (16-bit
floating point) support
_______________________________________________________
Follow-up Comment #1:
I would say this is a pretty low priority for Octave Developers. The
principal benefit is memory savings, but memory is really cheap these days (I
just checked Amazon and 8 GB DIMM module is approximately $30). The memory
savings are only likely to be apparent in systems with explicit hardware
support for this type such as FPGAs or ARM processors. That is pretty
esoteric and not something that most Octave users, with Intel or AMD CPUs, are
going to see. So, the benefits aren't that high, but the developer time would
be substantial. All of the basic mathematical functions (abs(), sin(), etc.)
would need to be extended to accept another input type besides single, double,
and the integer types. It can be done, just a lot of work.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?65046>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/