Re: Concertlina II: Full Circle

Liste des GroupesRevenir à c arch 
Sujet : Re: Concertlina II: Full Circle
De : mitchalsup (at) *nospam* aol.com (MitchAlsup1)
Groupes : comp.arch
Date : 24. Jun 2024, 00:46:23
Autres entêtes
Organisation : Rocksolid Light
Message-ID : <ae06566919e382be831d34e1363aa281@www.novabbs.org>
References : 1 2 3 4 5 6 7 8 9 10 11
User-Agent : Rocksolid Light
John Savard wrote:

On Sun, 23 Jun 2024 16:19:27 +0000, mitchalsup@aol.com (MitchAlsup1)
wrote:

While the vectorizing HW certainly has CRAY-like vector flip-flops
they are not addressable by SW. The code within the VEC--LOOP brackets reads as if scalar:: So, My 66000 consumes exactly 2
OpCodes to provide an entire vector instruction set--one that
works as well as possible across various implementations.

Oh, yes, your VVM is wonderful.
Well, at lest is keeps the R in RISC meaning Reduced instead of
Ridiculous.
It also alters cache semantics to avoid a single vector from erasing
the whole data cache. If the data is not going to be used again,
then it is not put in the cache (both inbound and outbound.)

My attempt at an imitation of VVM, at least, if not the real thing
that you have in your 66000, would be inferior in one important way to
Cray-style vector registers.

A virtual vector loop would take input vector values from memory, and
return results to memory. Yes, there are multiple operations within
the loop, but I am still assuming that the length and complexity of
such loops is constrained.

So if you have Cray-style vector registers, you have a place to store
intermediate results _between_ these loops that avoids referring to
memory.
Vector reduction is about the only realistic limitation, even here,
CRAY-like vectors "have their own problems". Performing a Summation
over an array consumes memory in order but performs FADDs in modulo
order whereas VVM performs the FADDs in program order. IEEE went so
far as to specify augmented addition which greatly ameliorates the
addition order problems.
These are loops from memory but to registers. About the only loops
that are from registers and to memory are memset()-like--which
easily vectorizes.
What VVM does not provide is the non-looping individual instructions.

In addition, one potentially catastrophic limitation is that, because
the meaning of register specifications in instructions is changed,
_there can't be any subroutine calls in such loops_. (Now that it's
typical for computers to have instructions that do log and trig
functions, this is slightly _less_ catastrophic, though.) Branches
within the loops and instruction predication, though, would still be
permitted.
First most trig functions have become instructions not subroutine
calls,
so that issue is ameliorated.
But, yes, VVM <as of now> only vectorizes the inner most loop.

John Savard

Date Sujet#  Auteur
17 Jun 24 * Concertlina II: Full Circle39John Savard
17 Jun 24 `* Re: Concertlina II: Full Circle38MitchAlsup1
17 Jun 24  `* Re: Concertlina II: Full Circle37John Savard
18 Jun 24   `* Re: Concertlina II: Full Circle36MitchAlsup1
18 Jun 24    +* Re: Concertlina II: Full Circle34John Savard
18 Jun 24    i+* Re: Concertlina II: Full Circle31John Savard
18 Jun 24    ii+* Re: Concertlina II: Full Circle9Thomas Koenig
18 Jun 24    iii`* Re: Concertlina II: Full Circle8John Savard
18 Jun 24    iii `* Re: Concertlina II: Full Circle7Thomas Koenig
18 Jun 24    iii  +* Re: Concertlina II: Full Circle2John Savard
19 Jun 24    iii  i`- Re: Concertlina II: Full Circle1John Savard
19 Jun 24    iii  `* Re: Concertlina II: Full Circle4Stephen Fuld
19 Jun 24    iii   +- Re: Concertlina II: Full Circle1MitchAlsup1
19 Jun 24    iii   +- Re: Concertlina II: Full Circle1MitchAlsup1
19 Jun 24    iii   `- Re: Concertlina II: Full Circle1MitchAlsup1
18 Jun 24    ii`* Re: Concertlina II: Full Circle21MitchAlsup1
18 Jun 24    ii `* Re: Concertlina II: Full Circle20John Savard
18 Jun 24    ii  +* Re: Concertlina II: Full Circle18MitchAlsup1
19 Jun 24    ii  i+* Re: Concertlina II: Full Circle16John Savard
19 Jun 24    ii  ii`* Re: Concertlina II: Full Circle15MitchAlsup1
19 Jun 24    ii  ii +* Re: Concertlina II: Full Circle8John Savard
20 Jun 24    ii  ii i`* Re: Concertlina II: Full Circle7John Savard
23 Jun 24    ii  ii i +* Re: Concertlina II: Full Circle2John Savard
25 Jun 24    ii  ii i i`- Re: Concertlina II: Full Circle1John Savard
23 Jun 24    ii  ii i `* Re: Concertlina II: Full Circle4MitchAlsup1
23 Jun 24    ii  ii i  `* Re: Concertlina II: Full Circle3John Savard
24 Jun 24    ii  ii i   `* Re: Concertlina II: Full Circle2MitchAlsup1
24 Jun 24    ii  ii i    `- Re: Concertlina II: Full Circle1John Savard
19 Jun 24    ii  ii `* Re: Concertlina II: Full Circle6Terje Mathisen
19 Jun 24    ii  ii  +* Re: Concertlina II: Full Circle2MitchAlsup1
19 Jun 24    ii  ii  i`- Re: Concertlina II: Full Circle1Terje Mathisen
19 Jun 24    ii  ii  `* Re: Concertlina II: Full Circle3Thomas Koenig
19 Jun 24    ii  ii   `* Re: Concertlina II: Full Circle2MitchAlsup1
19 Jun 24    ii  ii    `- Re: Concertlina II: Full Circle1Terje Mathisen
20 Jun 24    ii  i`- Re: Concertlina II: Full Circle1John Savard
19 Jun 24    ii  `- Re: Concertlina II: Full Circle1MitchAlsup1
18 Jun 24    i`* Re: Concertlina II: Full Circle2MitchAlsup1
18 Jun 24    i `- Re: Concertlina II: Full Circle1BGB-Alt
19 Jun 24    `- return for loop-back edges (Re: Concertlina II: Full Circle)1Anton Ertl

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal