9front - general discussion about 9front
 help / color / mirror / Atom feed
From: Kurt H Maier <khm@sciops.net>
To: 9front@9front.org
Subject: Re: [9front] 19,21 ARM Dev Summit and porting to Solid-Run hardware
Date: Sun, 10 Oct 2021 15:16:20 -0700	[thread overview]
Message-ID: <YWNmNM1vStzI65OE@wopr> (raw)
In-Reply-To: <1FEDF913F9FC55C72BF98FA67420A07B@9p.sdf.org>

On Sun, Oct 10, 2021 at 07:56:36PM +1100, v@9p.sdf.org wrote:
> Quoth Kurt H Maier <khm@sciops.net>:
> > I know a couple of
> > us have i.MXM8MQ devices, but ARM ports in general suck because there
> > is no standardization.
> 
> Pardon my French.  The MNT Reform (laptop of legacy standard parts) looks *shit*.

Not all of us are stupid enough to judge a laptop by what ports it has
built in.

> What you say about industry practices is true from your point of view.  Maybe the 99% is invisible to you and the people in the small niches you know of.  I understand the Fujitsu Fugaku Arm chip tops HPC for now.  I know Tesla uses Red Hat, NASA and others use AWS, and you can't say Tesla has not transformed autotech.

ARM can succeed in HPC because there are enough resources to custom-port
a linux distro to the specific board used in the cluster.  This is
feasible when there are tens of thousands of identical compute nodes,
as in the K computer, and not feasible when there are a shitload of
slightly-different disposable devices, which is the rest of the ARM
platform.  Pick a random ARM device in the wild, and maybe a few hundred
people are interested in running Linux on it.  Somewhere between one and
zero people want to run Plan 9.  There's just no return on that sweat
investment.

"NASA uses AWS" is not the same thing as "NASA uses AWS for HPC."
Everybody uses AWS and Azure for little things, because Microsoft and
Amazon are so desperate for clients that they'll give resources away
just for the opportunity.  But in the end, NASA still has supercomputers
on-prem in Goddard and Ames, and that's where the science happens.
Visualization webshit is a great candidate for AWS.

And I can absolutely say Tesla has not transformed auto tech.  Nobody
else is coming anywhere close to making the mistakes Tesla makes.  And
anyway, the recall from journald filling up the NAND flash was to
replace an Intel board.  Maybe they use ARM on the giant iPad glued to
the dashboard?  Maybe they can use that Coral dev board to do some
computer vision and finally do something about their massive panel gaps.

> For me a happy outcome would be to score resources from ARM and Solid-Run for 9front devs to work on, charming the marketing execs and their aspirated visions for the future we all want to have is fine by me.

ARM can provide nothing to 9front.  They're an IP holding company with
nothing useful to contribute to any computing project except, possibly,
money.  SolidRun is just another ARM OEM, suffering from the same
ridiculous problems that the rest of its market space does.  I don't
know what "resources" you think they can provide, but I hope you won't
ask for any on our behalf, because nothing good ever came of a corporate
sponsor.

khm


      parent reply	other threads:[~2021-10-11  9:08 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-10  5:28 v
2021-10-10  7:40 ` Kurt H Maier
2021-10-10  8:56   ` v
2021-10-10 14:53     ` hiro
2021-10-10 22:16     ` Kurt H Maier [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YWNmNM1vStzI65OE@wopr \
    --to=khm@sciops.net \
    --cc=9front@9front.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).