Message-ID:

To kick or not to kick... -- Somewhere on IRC, inspired by Shakespeare

devel / comp.arch / Re: Centrifuge instruction

Re: Centrifuge instruction

<09719ac5-ccee-45db-9c86-e7a34cc34ecen@googlegroups.com>

https://news.novabbs.org/devel/article-flat.php?id=32229&group=comp.arch#32229

X-Received: by 2002:ad4:5888:0:b0:61b:5bdd:d2b7 with SMTP id dz8-20020ad45888000000b0061b5bddd2b7mr5550431qvb.8.1684220222600;
Mon, 15 May 2023 23:57:02 -0700 (PDT)
X-Received: by 2002:a05:6870:108f:b0:195:af27:8e94 with SMTP id
15-20020a056870108f00b00195af278e94mr14091507oaq.8.1684220222318; Mon, 15 May
2023 23:57:02 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder6.news.weretis.net!news.misty.com!border-2.nntp.ord.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.arch
Date: Mon, 15 May 2023 23:57:02 -0700 (PDT)
In-Reply-To: <8c97deb8-bbeb-4c9d-9883-c527bc30bac9n@googlegroups.com>
Injection-Info: google-groups.googlegroups.com; posting-host=99.251.79.92; posting-account=QId4bgoAAABV4s50talpu-qMcPp519Eb
NNTP-Posting-Host: 99.251.79.92
References: <u2016l$9b5v$1@newsreader4.netcologne.de> <2023Apr27.080111@mips.complang.tuwien.ac.at>
<u2db12$1r936$1@dont-email.me> <dace43b5-9d12-48c5-90ad-4f26add1b15en@googlegroups.com>
<u2gg61$2emt2$1@dont-email.me> <u2grh3$2gp8v$1@dont-email.me>
<zVS2M.506047$cKvc.496132@fx42.iad> <u2hjbj$2kg3l$1@dont-email.me>
<u2jdvn$30fef$1@dont-email.me> <u2mhfh$3mja7$1@dont-email.me>
<u2o3hc$44ri$1@dont-email.me> <2023May1.163832@mips.complang.tuwien.ac.at>
<M9S3M.582961$5CY7.211116@fx46.iad> <2023May1.214819@mips.complang.tuwien.ac.at>
<985a755e-e984-4011-9f44-ae8e2061c347n@googlegroups.com> <d0ec4e5b-6cc4-4d9d-9770-5d2f123e9f9dn@googlegroups.com>
<285c7b6d-2e8b-432a-8b91-3d8630d6695dn@googlegroups.com> <2682dc68-f68a-4527-b3fa-363d06f6cf29n@googlegroups.com>
<604df68d-dbd9-4c53-bb7f-be2cf1e0fe45n@googlegroups.com> <66e20780-90f0-4d99-b4ca-7158ae3e8676n@googlegroups.com>
<27537a42-d36c-4241-8856-a784b389e442n@googlegroups.com> <2d8e2487-c184-4c4a-b729-5e9b43d9e5b7n@googlegroups.com>
<e31ef5c0-f384-4eb1-98c1-a4191cd893b8n@googlegroups.com> <c25e98cb-eea9-42f9-8300-6e61067ae100n@googlegroups.com>
<4244ebfd-4724-463f-8ff6-47a256724704n@googlegroups.com> <8c97deb8-bbeb-4c9d-9883-c527bc30bac9n@googlegroups.com>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <09719ac5-ccee-45db-9c86-e7a34cc34ecen@googlegroups.com>
Subject: Re: Centrifuge instruction
From: robfi680@gmail.com (robf...@gmail.com)
Injection-Date: Tue, 16 May 2023 06:57:02 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Lines: 34

by: robf...@gmail.com - Tue, 16 May 2023 06:57 UTC

On Tuesday, May 16, 2023 at 12:13:33 AM UTC-4, Quadibloc wrote:
> On Wednesday, May 10, 2023 at 9:54:54 AM UTC-6, MitchAlsup wrote:
>
> > Ah,
> > <
> > It is not that you are assuming the problem is larger than it is.
> > <
> > It is that you are making the problem larger than it is.
> > <
> > Gotcha.
> >
> Yes, I may be making the problem larger than it has to be.
>
> But it's still not all that large, even so - and, since I admit that the block
> structure is a barrier to acceptance, I think I am going to investigate the
> possibility of going to very simple variable-length instructions.
>
> John Savard

For Thor2024 going with variable length instructions in 32-bit parcel
increments, like several other designs. The length of the instruction is
determined by the first eight bits. I tried to save over having bits dedicated
to the instruction length since most instructions are 32-bit. Since most
instructions are 32-bit the table is small, lots of duplicate entries. But
having the code byte align-able. The idea is to allow potentially shorter
instructions and to allow code to be positioned at any byte which may
make it more difficult to hack. Also having a code scrambler in the front
end, which is just an XOR to a random value. It should make the program
look different.

Subject	Author
Re: Centrifuge instruction	robf...@gmail.com