Did you try swapoff -a on your Raspi 400? Did it work remarkable faster
oder did I suffer from imagination?

FW

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On Wed, 4 May 2022 08:01:27 +0200, F. W. wrote:

Did you try swapoff -a on your Raspi 400? Did it work remarkable faster
oder did I suffer from imagination?

FW

If you meant me; I did not.

Where would I put that ?

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

Am 04.05.2022 um 09:15 schrieb GrailKing:

On Wed, 4 May 2022 08:01:27 +0200, F. W. wrote:

Did you try swapoff -a on your Raspi 400? Did it work remarkable
faster oder did I suffer from imagination?

FW

If you meant me; I did not.

Where would I put that ?

I meant everybody (who owns a 400).

In a commandline I did sudo swapoff -a and think my 400 is a little bit
faster. Can be that my SD-card is very slow.

FW

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On Wed, 4 May 2022 13:24:38 +0200, F. W. wrote:

Am 04.05.2022 um 09:15 schrieb GrailKing:

On Wed, 4 May 2022 08:01:27 +0200, F. W. wrote:

Did you try swapoff -a on your Raspi 400? Did it work remarkable
faster oder did I suffer from imagination?

FW

If you meant me; I did not.

Where would I put that ?

I meant everybody (who owns a 400).

In a commandline I did sudo swapoff -a and think my 400 is a little bit faster. Can be that my SD-card is very slow.

FW

try
Free -h so see how much memory is used
the amount of swap used on any of mi Pi's is negligible. my deductions
from this are:
If it is not being used then it cant be slowing the system down.
It is probably not needed anyway.

this will probably vary depending on your own usage patterns






--
Repel them. Repel them. Induce them to relinquish the spheroid.
-- Indiana University football cheer

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On Wed, 4 May 2022 18:14:53 -0000 (UTC), alister wrote:

try Free -h so see how much memory is used the amount of swap used on

You can also see this information on the 'top' display along with other
useful stuff about what process(es) are hogging memory and/or CPU time and
how much memory is being occupied by file buffering (often more than is occupied by executable code).






--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On 04/05/2022 12:24, F. W. wrote:

In a commandline I did sudo swapoff -a and think my 400 is a little bit faster. Can be that my SD-card is very slow.

If you aren't using much memory, it wont be any faster. If you are using
too much memory swapping to an SD card will cause it to run very slowly,
but not having any swap will cause it to hang while it attempts to kill
off processes, or if you are unlucky hang altogether.

---druck

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On 04/05/2022 13:24, F. W. wrote:

I meant everybody (who owns a 400).

In a commandline I did sudo swapoff -a and think my 400 is a little bit faster. Can be that my SD-card is very slow.

SD cards are quite slow, yes. If you need a lot of swap, then swap to a
faster device. I used to have swap on a ZFS RAIDZ array spread across 4
× 2TB USB3 drives on a USB3 hub, and it seemed pretty fast to me.

On my older Pi 3 or Pi 3B, I configured ZRAM for in-memory compressed
swap, which is what Ubuntu is now doing by default in version 22.04 for
the Pi. I was doing it in 14.04 and it worked pretty well: Lubuntu with
LXDE and ZRAM was quite snappy for such a low-end machine.

But as druck says: just turning it off completely is not a good move and
will have negative side-effects.

A small USB3 SSD will only cost about as much as a 4 or 8GB Pi 4 itself
and will substantially improve performance. If you don't want to
experiment with USB booting, you can leave your /boot partition on the
SD card, and just put the root and swap partitions on the USB3 drive.


--
Liam Proven ~ Prague, Czechia
[email protected]
(or liamproven on either AOL or Yahoo UK)

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On Wed, 4 May 2022 21:23:19 -0000 (UTC), Martin Gregorie wrote:

On Wed, 4 May 2022 18:14:53 -0000 (UTC), alister wrote:

try Free -h so see how much memory is used the amount of swap used on

You can also see this information on the 'top' display along with other useful stuff about what process(es) are hogging memory and/or CPU time
and how much memory is being occupied by file buffering (often more than
is occupied by executable code).

Linux is designed to use spare memory as buffers whenever possible,
it is released if it is needed for other purposes.




--
Using TSO is like kicking a dead whale down the beach.
-- S. C. Johnson

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)

On Thu, 5 May 2022 18:47:48 -0000 (UTC), alister wrote:

On Wed, 4 May 2022 21:23:19 -0000 (UTC), Martin Gregorie wrote:

On Wed, 4 May 2022 18:14:53 -0000 (UTC), alister wrote:

try Free -h so see how much memory is used the amount of swap used on

You can also see this information on the 'top' display along with other
useful stuff about what process(es) are hogging memory and/or CPU time
and how much memory is being occupied by file buffering (often more
than is occupied by executable code).

Linux is designed to use spare memory as buffers whenever possible,
it is released if it is needed for other purposes.

Of course: I thought that was worth pointing out because it can be a
surprise to a Linux newbie. Its also where the speed comes from, and is particularly noticeable when writing code in C or Java on a slowish
machine with a decent amount of RAM, say an i5-based laptop with 8GB RAM.

The first compilation in a code writing session can be very much faster
than subsequent recompiles because the first one needs to load make, the C compiler and linker (or the JVM, ant and the Java compiler) as well as the source code, while subsequent compiler runs will be anything up to 400%
faster because the compiler chain binaries and the source files you're
working on will still be in RAM from the previous compilation run.

I've timed it for Java: one fairly large project (a duty roster building
visual application for a sport club took noticeably longer for the first compilation of the day. I got intriqued, so I measured it to be sure it
wasn't a subjective effect by running the compilation under 'time'. It was real: the first compilation of the day took 4 seconds and subsequent compilations took less than a second.

That said, I don't see the same speedup effect on any Pi I use, but I
don't have a Pi 4, but I would expect to see it on a Pi 4 that is maxed-
out on RAM.


--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: Agency HUB, Dunedin - New Zealand | Fido<>Usenet Gateway (3:770/3)