[Libre-soc-dev] results of phone call about simple-v prefix

[Luke Kenneth Casson Leighton]

On 11/27/20, Jacob Lifshay <programmerjake at gmail.com> wrote:

> not change it for the rest of the shader. Video stuff would probably not
> need to constantly change VL as well, since the most common operations
> would work on large blocks (8x8 or bigger) leaving plenty of time to change
> VL since the processor is multi-issue. We do need to double check with
> those working on Video.

that will be Lauri.

needs bugreport for that one.

also check with IBM: reservations of v3.1 64bit spare space.  another bugreport.

also idea resolving the tugowar on predication: support both integer
*and* CR predicate sources because each has advantages in different
circumstances.

this led to idea of having a new instruction similar to branch logic
eq/gt/lt that has better coverage of CR bits 0-3 and alleviates need
for crand/cror as is currently the case with branch.

when vectorised such an instruction would create the onebit
needed/used by predication.

one objection to that which may come up is, "use macro-op fusion" which, hmmm...

anyway a new bugreport for that, too.

l.