On Mon, Mar 16, 2026 at 12:42:47PM +0100, qwx via 9fans wrote:
> On Mon Mar 16 12:11:34 +0100 2026, [email protected] wrote:
> > I'm continuously trying to improve arguments to sell the idea that
> > Plan9 has its points.
> >
> > The present context (and enduring context) of explosion of energy
> > costs is an opportyunity to leverage.
> >
> > Are there some studies comparing the energy costs between OSes?
> > demonstrating that a simpler and smaller OS can not only run on a low
> > energy device (that can't host a fat OS) and that, furthermore, even
> > on the same hardware, for the same task, plan9 uses less resources
> > than another OS?
> >
> > I'm convinced, for example, that if "researchers" about the so-called
> > "global warming" were redacting their papers using kerTeX, on Plan9,
> > instead of TexLive, we will not be in a "global warming" era but
> > in a glacial one i.e. that this is the "global warming" buzz and
> > the exchange and the papers around it that are wasting energy and
> > generating the denounced heat.
>
> If your argument is about power consumption, let's keep in mind the
> fact that we have practically no powersaving code whatsoever for any
> hardware and typically choose to ignore the problem. We don't even
> have suspend to RAM or disk. Our model also almost mandates running
> multiple machines with no interruption as part of our distributed
> network. On the other hand, it's possible to install a linux
> distribution to any Plan 9-supported hardware and configure it in a
> way that aggressively saves power. I went pretty far down the rabbit
> hole at one point getting a x220 thinkpad or similar to use less than
> 5W most of the time. No doubt compiling a Latex document will spike
> consumption, but either way I don't think energy savings is where
> Plan 9 wins.
>
Yes, but in my view, if the hardware is mainly idle, it is because
it is not correctly profiled for the task (the best energy
saving is to shut down the unused hardware...). So it could be
argued that the advantage of the powersaving code is here to
compensate for an original waste of resources---this is not to say
that powersaving features are useless, since for a terminal it is not
unusual to have not negligible periods of "do nothing"; but when trying
to judge the pros and cons, no initial configuration has to be taken as
an undisputed fact.
The main energy waste is, today, the GPU (you need to have a mean to
display at least text, for interaction, but even the not discrete GPU
on SoC are overkill for just that). Furthermore, with a simpler
OS, one could build dedicated hardware knowing that customizing a
kernel for such special hardware is doable (hardware optimized for a
kind of task instead of a general purpose chip; this is already in part
what RISC-V is describing, segregating group of features in the
instruction set).
--
Thierry Laronde <tlaronde +AT+ kergis +dot+ com>
http://www.kergis.com/
http://kertex.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C
------------------------------------------
9fans: 9fans
Permalink:
https://9fans.topicbox.com/groups/9fans/Tacc477007dd06ec0-Mc1c4ca6904dca170685cd081
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
