[libre-riscv-dev] evaluation of bit-witdh scoreboards

Luke Kenneth Casson Leighton lkcl at lkcl.net
Mon Dec 17 10:54:32 GMT 2018


---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Mon, Dec 17, 2018 at 10:21 AM Jacob Lifshay <programmerjake at gmail.com> wrote:
>
> On Mon, Dec 17, 2018, 02:12 Luke Kenneth Casson Leighton <lkcl at lkcl.net
> wrote:
>
> > Still seems like it is worth an op, to do 1 FPx4 to RGB565 per cycle.
> >
> Yeah, we also need one for R5G5B5A1 since that's another common 16-bit
> format.

 ok.

> >
> > Hm can we use FP16 on the pixel colours? Might be able to get 2x speed that
> > way.
> >
> I don't think so, since the types the Vulkan spec specifies are 32-bit
> floats.

 ohhh and it's an abstracted API, damn.  so we'd need to rewrite the
Vulkan layer (or propose the addition as part of the spec).

l.



More information about the libre-riscv-dev mailing list