Ask stupid questions and you'll get a stupid answers, was: Technological advantages over Linux

[Polytropon]

On Sat, 25 Jul 2020 22:46:23 -0400, Aryeh Friedman wrote:
> On Sat, Jul 25, 2020 at 9:06 PM Valeri Galtsev <galtsev at kicp.uchicago.edu>
> wrote:
> [...]
> >
> > Maybe you have some piece of software with memory leak (Oh, I just
> > noticed you have said it yourself), and big enough swap space (like
> > recommended 4 GB), and your machine starts swapping a lot. That can
> > bring any machine to its knees. Until swap is exhausted, and system
> > kills the offender.
> >
> 
> Firefox, a terminal and a audio player (deadbeef) and thats it.   12
> logical cores [AMD Ryzen 5 2600] 24 GB of RAM with 6 GB of swap backed by a
> 1 TB SSD.  The only thing I can think of is there is a HTML 5 game that I
> often leave open all the time (one of those hurry up and wait multiplayer
> strategy/war freemium games [Forge of Empires]) and that slows it down but
> almost anything else I do besides the above also slows it down (for example
> libreoffice slows the entire machine to a crawl if you select a cell range
> in calc for copy/pasting [you don't need to actually copy and paste
> anything just select the region]).

Interesting. Cannot confirm on far inferior hardware (10 years
old Lenovo laptop, 4 GB RAM, 300 GB SSD). :-)



> If I start mysql-server it makes the
> machine go out to lunch if I am using the GUI at all. 
> [...]
> So the only possible conclusion I can draw
> is it is X being extremely screwy on FreeBSD due to all the linux hacks in
> it (this despite X's claim it will work on any POSIX machine with the right
> video support).   As we move more and more towards wayland as being the
> pixel driver it gets worse.

It hasn't alwas been the case. Even if we can assume that
program complexity has increased, and so have hardware demands,
compared to how memory and computing power have increased,
my general observation is that "overall usage speed" has
dropped. I don't say software has gotten slower (probably
it has), but hardware didn't compensate it so far. Certain
things are much faster, especially dedicated stuff like
media decoding and encoding, or 3D rendering, or I/O, or
network, but what you see, as a user, often hasn't.

I have no valid explanation.

When I compare today's environments where users complain
about skipping audio when they play a MP3 file and move some
windows across the screen, I cannot even imagine what I've
been doing wrong for decades - because I never had that kind
of problem. One of my first FreeBSD workstations was a 150 MHz
Pentium PC with 32 MB RAM (that's Mega, not Giga!), later on
upgraded to 128 MB, with a 10 GB PATA ("IDE") hard disk. On
that system, I was able to play MP3 while downloading a
FreeBSD installation CD, at the same time having compile
some port, burn a previously downloaded CD, and browse the
web with a still responsive Opera web browser (which, at
that time, was the Linux version). Do you thin this sounds
wrong? ;-)



> > Incidentally, I've got question for experts. How "Out Of Memory"
> > situation is handled by FreeBSD. I only dealt with [memory] "leaky"
> > things under Linux, and OOM killer there really does its job well (and
> > its way of judgement I do know).
> >
> 
> When out of swap and real memory it just starts killing of processes that
> have not made any CPU requests recently.

I tried to experiment with that on my old home PC, a real
el-cheapo supermarket PC from 2007: In the web browser, I
opened about 70 tabs with media content playing. I got this:

  1  [||||                      7.9%]     Tasks: 73, 0 thr; 1 running
  2  [|                         2.6%]     Load average: 0.38 0.70 0.42 
  Mem[|||||||||||||||||||1601/1990MB]     Uptime: 00:52:33
  Swp[|||||               275/2047MB]

CPU usage was increasing from time to time, and swap usage was
also growing:

  Mem[|||||||||||||||||||1597/1990MB]
  Swp[||||||              404/2047MB]

The system was quite slow because of the swap I/O, but it was
still responding. I could close several tabs and return the
system to "halfway normal".

Again, this sounds totally strange.



> > >       Yeuch, but it has nothing to do with making X work, X works just
> > > fine for as long as you care to leave it running. I can't speak for Gnome
> > > or KDE I don't use them, but with a simple WM X is as rock solid as it
> > has
> > > ever been.
> >
> >
> The above is true regardless of WM or desktop, it is true of all Linux
> derived GUI stuff.
> 
> TL; DR -- X and wayland even more so have become far too tied to Linux for
> them to really be considered workable on any other OS.

There is some truth in this statement. Their ports available
on FreeBSD compile, work, and run more or less like on Linux,
but due to the fact that FreeBSD != Linux, there are some gaps
in functionality. Workarounds created with piles of abstraction
libraries don't help much.



-- 
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...