a man crying alone in the wilderness

"And, now for an entirely different circuit ..."

While 'tinkering' with Mr Moore's 25X model,

I modified my basic model , some while back,
( URL,
http://groups.google.com/group/comp....00007a63ce9072

)
to develop an any 16-channel to any 16-channel switch model,
( modified by DSPs for an SMP feature, along diagonal) that

can be constructed by daisy chaining four
( 4) 25X models, ( Mr. Moore's 25X model)

HOWEVER, intra-switch messaging times increase
( DSPa[1..16] <--> DSPb[1..16]) and

result in reduced SMP messaging performance and
some MPP channel to channel messaging are anomolous
with my original ( URL, above) basic model and VLIW theory.


16-way by 16-way >= 256 elements,

VS

16-way by 16-way >= 100 elements, ( four 25X models, /daisy chain/
connected thru their fifth channel )

---

URL,
http://groups.google.com/groups?enc_...3Zgw&scoring=d

Oh, Goodness me!

---