[Libre-soc-dev] new svp64 page
programmerjake at gmail.com
Thu Dec 10 03:35:47 GMT 2020
On Wed, Dec 9, 2020 at 6:45 PM Luke Kenneth Casson Leighton
<lkcl at lkcl.net> wrote:
> i don't entirely know why we haven't yet raised this one as a
> bugreport :) perhaps i just like the different perspective.
> in looking at the AV opcodes i realised that DSP Audio saturation (if
> sum greater than max int range set to 0xffff) would either need a ton
> of ops (yuk) or an SPR bit (yuk) or, much better, a spare bit or two
> in the 24 bits from the SV Prefix
Couldn't we just use the already existing VSX instructions for some of
that and just SV-ify them, where the SV prefix just takes the base
using a SV-prefixed form of vaddshs would act as if the vaddshs was
just a scalar signed 16-bit saturating add, adding SV-style
vectorization if the src/dest registers are set as vectors.
using a SV-prefixed form of vavguw would act as if the vavguw was just
a scalar unsigned 16-bit average operation.
Using those (and other similar instructions) would make it waay easier
to get the operations we need with less friction, since the encodings
More information about the Libre-soc-dev