[Libre-soc-dev] Triangular REMAP schedule

[lkcl]

without going overboard into ZOLC territory, after jacob and i worked on bigint last month i have been thinking how to do the outer loops of Knuth algorithms D and M, at least in Vertical-First Mode.

it would be a fascinating application of SVP64 in that for the first time some of SVSTATE would need to be context-switched, i.e. the computation of a partial intermediary result within an outer loop would roll SVSTATE forward, and would need rolling back before calculating the next stage.

for multiply there is a register offset that is in effect (x+y)
when assuming x is the element from src1 and y from src2

....0000
...1111.
..2222..
.3333...

this should be dead easy to calculate these kinds of offsets: i would like to make it more general, and include simple triangular mapping so as not to have to add conflictd.

are there any other kinds of regular algorithm patterns up to 3 nested loops anyone can think of?

l.