Path: ...!news.nobody.at!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: vallor <vallor@cultnix.org>
Newsgroups: comp.os.linux.advocacy
Subject: Two points
Date: 29 Nov 2024 14:52:53 GMT
Lines: 32
Message-ID: <lqu2q4Fa9h6U1@mid.individual.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
X-Trace: individual.net yig8fqnd8uMkoB2I29bgzg0KPRm9GSL81toIC4WkMham58nDCF
Cancel-Lock: sha1:+FnkGrtffcXizLYw+UIecVa4MTA= sha256:AlNUnRh+ZZVYyQrxQz/KUz2LmquK4sYK/QYJheDEUDQ=
X-Face: +McU)#<-H?9lTb(Th!zR`EpVrp<0)1p5CmPu.kOscy8LRp_\u`:tW;dxPo./(fCl
 CaKku`)]}.V/"6rISCIDP`
User-Agent: Pan/0.161 (Hmm2; be402cc9; Linux-6.11.10)
Bytes: 1986

Two points:

1) It's surprisingly easy to get the Hurd running in a
virtual machine with kvm:

$ uname -a
GNU hurd 0.9 GNU-Mach 1.8+git20240714-up-amd64/Hurd-0.9 x86_64 GNU

2) I own two Mac mini's, which are sitting in a drawer.  They
were made from notebook equipment, and they are crap.

One of them came with 4GB of RAM, and ran like a dog.  Thankfully,
I could add more (notebook) RAM, bringing it up to 16GB.  Then
it ran acceptably, probably because it was caching the slow, slow
drive it came with.

So we have a Mac Studio now, which is a low-end UNIX workstation.
They aren't cheap, with a price on the order of magnitude of what
I paid for my Linux workstation.

Eyeballing the results, it looks like it renders Fooocus images about 1/4
the speed of my Linux workstation.  I thought it would be a lot
faster, but I guess pytorch doesn't support hw acceleration yet -- or
the hardware on the Mac is no match for an RTX 3090ti w/24G of memory.

You can say what you want, but I'd much rather use my Thelio
than the Mac Studio.  YMMV.

-- 
-v System76 Thelio Mega v1.1 x86_64 NVIDIA RTX 3090 Ti
   OS: Linux 6.11.10 Release: Mint 21.3 Mem: 258G
   "Mary had a little RAM -- only about a MEG or so."