The Natural Philosopher <tnp@invalid.invalid> wrote:
> On 03/05/2024 00:23, Computer Nerd Kev wrote:
>> I run current Firefox on a PC with 2GB RAM
>> and I don't have it getting killed by the kernel, nor do I have
>> problems with kernel crashes/reboots. I've also tried running
>> recent Firefox on a PC with 512MB RAM and noticed that it performs
>> much worse than with 2GB RAM, slowing down to a crawl while loading
>> some websites, suggesting that it really does need more RAM in that
>> case.
>>
>> I don't use web browsers to play video. If you're streaming super
>> high resolution video through your browser with the latest and
>> greatest compression algorithms, then it probably has the right
>> to chew up a lot of RAM.
>>
> I use browsers to play videos. RAM usage is not massive. CPU usage is.

Perhaps RAM usage isn't that great on the scale of systems with tens
of GB of RAM. Compared to 2GB I think it may be significant because
when I tried to transcode a 1920x1080 video in AV1 format to
lower-res MPEG4 on a VPS with 1GB RAM + 1GB swap space, ffmpeg
(v. 4) surprised me by running out of RAM. Of course it could be a
bug in that decoder, but not having tackled much AV1 video before I
concluded that high resolutions plus modern video codecs requires
lots of RAM.

> What chews memory are *commercial* websites loaded with (deliberately)
> buggy javaScript that cause javaScript engines to go into meltdown. How
> that is handled is browser dependent.
>
> Ublock Origin helps massively, but is not a complete answer.
> What is a massive help at leats in Mint Mate is the System Monitor
> widget that I keep in the task bar permanently displaying CPU, RAM and
> Network usage as teeny graphs.
> It is very easy to see when memory is all grabbed by a process rather
> than simply cache
> And monitor how much gets released when you close a website page.

There's also about:performance in Firefox. I do use NoScript with
all my browsing in Firefox.

--
__ __
#_ < |\| |< _#

On 03/05/2024 06:51, The Natural Philosopher wrote:
> On 02/05/2024 15:01, James Harris wrote:
>> On 01/05/2024 21:42, Kenny McCormack wrote:
>>> In article <v0tkdc$3780t$1@dont-email.me>,
>>> Lew Pitcher  <lew.pitcher@digitalfreehold.ca> wrote:
>>>> On Wed, 01 May 2024 11:32:18 +0100, James Harris wrote:
>>>>
>>>>> Not a question, just an observation.
>>> ...
>>>>
>>>> So, if you want help, how about providing some details.
>>>
>>> OP was pretty clear that he was *not* seeking help.
>>
>> Correct. I don't want to overtax the expert assistance that's clearly
>> available from posters to this newsgroup and would rather ask for help
>> relatively rarely, so I am not looking (at least, yet!) for a
>> solution. My point, in this thread, was, instead, to comment on what I
>> perceive to be an OS-design issue - a long-term interest.
>>
>>
> In short anti-linux trolling.

Nope. Posting to a discussion forum to have a discussion about a point
of interest.

--
James Harris

James Harris wrote:

>It may not be a memory leak but it's certainly a rogue program -
>chrome/chromium in this case, though I've had similar and worse problems
>with Firefox which seems to me more likely to have a genuine memory leak.

I had the issue with Chrome (not with Chromium) that when streaming
videos it would take as much RAM as the video size was in total, so one
browser tab could easily eat up a couple of GB of memory. I then
switched to Chromium, which didn't have this issue. This was probably 6
months ago, so I do not know whether current versions of Chrome still
have his issue (Ubuntu here).

-jw-
--
And now for something completely different...

On 2024-05-03 01:23, Computer Nerd Kev wrote:
> Carlos E.R. <robin_listas@es.invalid> wrote:
>> On 2024-05-02 16:06, James Harris wrote:
>>>
>>> Yes, I am (at least tentatively) blaming Linux but not for the problem
>>> you think. I wouldn't expect Linux to prevent programs from gobbling up
>>> memory but I would expect it to manage memory hogs more gracefully than
>>> it does. IME Windows handles the same situation better - and that may be
>>> down to the different designs of their paging systems.
>>>
>>> Don't get me wrong. I much prefer Linux to Windows and have often seen
>>> Windows get into a worse situation under different circumstances. But
>>> for page management ISTM that the design of Linux's paging subsystem may
>>> not be the best.
>>
>> You can configure Linux to crash the application that is behaving badly
>> by grabbing all the memory. It is up to you, the boss.
>>
>> The philosophy is not to nanny care for you. It does what you asked.
>> More memory? Yes sir. Till death does us part. Your orders will be obeyed.
>>
>> :-)
>
> I didn't know there was an option not to have it do that (via the
> Out Of Memory Reaper). It seems the alternatives are to reboot or
> risk the kernel crashing, so your description seems about right.
>
> https://www.oracle.com/technical-resources/articles/it-infrastructure/dev-oom-killer.html
>
> The OP has noted now that the process that consumes their RAM is
> Chrome or Firefox. I've not seen a detailed description of why it
> happens, but I've long noted that Firefox seems to expand its RAM
> usage to the available space, but different from a memory leak in
> that it usually leaves a certain amount free. I assume that this
> in intended behaviour. I run current Firefox on a PC with 2GB RAM
> and I don't have it getting killed by the kernel, nor do I have
> problems with kernel crashes/reboots. I've also tried running
> recent Firefox on a PC with 512MB RAM and noticed that it performs
> much worse than with 2GB RAM, slowing down to a crawl while loading
> some websites, suggesting that it really does need more RAM in that
> case.

Firefox uses a lot of memory.

One problem is memory fragmentation. It reserves many chunks then
eventually frees many, but they may not being contiguous, so maybe they
can not be reused, so it requests more memory.

When I started using Linux, back in 1998, I did notice that Linux needed
more memory than Windows to work right. I had to improve my hardware
because of that. My previous computer had 8 GB and it was not enough, it
was swapping. So now I have a machine with 64.

>
> I don't use web browsers to play video.

No youtube? :-)

> If you're streaming super
> high resolution video through your browser with the latest and
> greatest compression algorithms, then it probably has the right
> to chew up a lot of RAM.
>

--
Cheers, Carlos.

On 03/05/2024 08:45, Computer Nerd Kev wrote:
> The Natural Philosopher <tnp@invalid.invalid> wrote:
>> On 03/05/2024 00:23, Computer Nerd Kev wrote:
>>> I run current Firefox on a PC with 2GB RAM
>>> and I don't have it getting killed by the kernel, nor do I have
>>> problems with kernel crashes/reboots. I've also tried running
>>> recent Firefox on a PC with 512MB RAM and noticed that it performs
>>> much worse than with 2GB RAM, slowing down to a crawl while loading
>>> some websites, suggesting that it really does need more RAM in that
>>> case.
>>>
>>> I don't use web browsers to play video. If you're streaming super
>>> high resolution video through your browser with the latest and
>>> greatest compression algorithms, then it probably has the right
>>> to chew up a lot of RAM.
>>>
>> I use browsers to play videos. RAM usage is not massive. CPU usage is.
>
> Perhaps RAM usage isn't that great on the scale of systems with tens
> of GB of RAM. Compared to 2GB I think it may be significant because
> when I tried to transcode a 1920x1080 video in AV1 format to
> lower-res MPEG4 on a VPS with 1GB RAM + 1GB swap space, ffmpeg
> (v. 4) surprised me by running out of RAM. Of course it could be a
> bug in that decoder, but not having tackled much AV1 video before I
> concluded that high resolutions plus modern video codecs requires
> lots of RAM.
>
Transcoding is not playing

>> What chews memory are *commercial* websites loaded with (deliberately)
>> buggy javaScript that cause javaScript engines to go into meltdown. How
>> that is handled is browser dependent.
>>
>> Ublock Origin helps massively, but is not a complete answer.
>> What is a massive help at leats in Mint Mate is the System Monitor
>> widget that I keep in the task bar permanently displaying CPU, RAM and
>> Network usage as teeny graphs.
>> It is very easy to see when memory is all grabbed by a process rather
>> than simply cache
>> And monitor how much gets released when you close a website page.
>
> There's also about:performance in Firefox. I do use NoScript with
> all my browsing in Firefox.

I haven't yet needed to

>

--
Renewable energy: Expensive solutions that don't work to a problem that
doesn't exist instituted by self legalising protection rackets that
don't protect, masquerading as public servants who don't serve the public.

James Harris <james.harris.1@gmail.com> writes:
> On 03/05/2024 06:51, The Natural Philosopher wrote:
>> In short anti-linux trolling.
>
> Nope. Posting to a discussion forum to have a discussion about a point
> of interest.

There’s not really enough in the original posting for anyone to have an
informed discussion, though.

--
https://www.greenend.org.uk/rjk/

On 2024-05-03, Lawrence D'Oliveiro <ldo@nz.invalid> wrote:

> On Thu, 2 May 2024 14:47:24 +0100, James Harris wrote:
>
>> A number of years ago
>> I came up with what I thought was a sensible design for a paging system.
>> I later found that it corresponded very closely to the one used in
>> Windows and differed markedly from the paging system used in Linux. I
>> remember thinking that the latter was not a good design ...
>
> Look at all the effort those smart Microsoft engineers had to go to,
> just to get Windows file copying to work properly:
> <https://techcommunity.microsoft.com/t5/windows-blog-archive/inside-vista-sp1-file-copy-improvements/ba-p/723622>

Sounds like a lot of gratuitous complexity, aggravated by their penchant
for lying about when something is done. But M$ file copying issues are
nothing new - many of us still remember being stung by the MS-DOS COPY
command's refusal to copy a zero-length file (although it still deleted
a like-named file at the destination if it existed).

--
/~\ Charlie Gibbs | Microsoft is a dictatorship.
\ / <cgibbs@kltpzyxm.invalid> | Apple is a cult.
X I'm really at ac.dekanfrus | Linux is anarchy.
/ \ if you read it the right way. | Pick your poison.

On Fri, 03 May 2024 16:23:02 +0100, Richard Kettlewell wrote:

> There’s not really enough in the original posting for anyone to have an
> informed discussion, though.

Which is the point of trolling, isn’t it?

The Natural Philosopher <tnp@invalid.invalid> wrote:
> On 03/05/2024 08:45, Computer Nerd Kev wrote:
>> The Natural Philosopher <tnp@invalid.invalid> wrote:
>>> On 03/05/2024 00:23, Computer Nerd Kev wrote:
>>>> I don't use web browsers to play video. If you're streaming super
>>>> high resolution video through your browser with the latest and
>>>> greatest compression algorithms, then it probably has the right
>>>> to chew up a lot of RAM.
>>>>
>>> I use browsers to play videos. RAM usage is not massive. CPU usage is.
>>
>> Perhaps RAM usage isn't that great on the scale of systems with tens
>> of GB of RAM. Compared to 2GB I think it may be significant because
>> when I tried to transcode a 1920x1080 video in AV1 format to
>> lower-res MPEG4 on a VPS with 1GB RAM + 1GB swap space, ffmpeg
>> (v. 4) surprised me by running out of RAM. Of course it could be a
>> bug in that decoder, but not having tackled much AV1 video before I
>> concluded that high resolutions plus modern video codecs requires
>> lots of RAM.
>>
> Transcoding is not playing

They both involve decoding. I converted other HD video to the same
output format without RAM issues, so the difference was the input
format, decoding the AV1 video, which applies to playback and
transcoding. Maybe it was a bug in ffmpeg, but as I haven't seen
benchmarks for memory usage online so I concluded from my own
experience. After a recent OS upgrade the VPS now has ffmpeg 5
installed, so I'll see what happens next time some AV1 video comes
my way.

--
__ __
#_ < |\| |< _#

Carlos E.R. <robin_listas@es.invalid> wrote:
> On 2024-05-03 01:23, Computer Nerd Kev wrote:
>>
>> The OP has noted now that the process that consumes their RAM is
>> Chrome or Firefox. I've not seen a detailed description of why it
>> happens, but I've long noted that Firefox seems to expand its RAM
>> usage to the available space, but different from a memory leak in
>> that it usually leaves a certain amount free. I assume that this
>> in intended behaviour. I run current Firefox on a PC with 2GB RAM
>> and I don't have it getting killed by the kernel, nor do I have
>> problems with kernel crashes/reboots. I've also tried running
>> recent Firefox on a PC with 512MB RAM and noticed that it performs
>> much worse than with 2GB RAM, slowing down to a crawl while loading
>> some websites, suggesting that it really does need more RAM in that
>> case.
>
> Firefox uses a lot of memory.
>
> One problem is memory fragmentation. It reserves many chunks then
> eventually frees many, but they may not being contiguous, so maybe they
> can not be reused, so it requests more memory.

It seems to me like it looks at the available RAM (not including
swap space) and only works at freeing space when the free RAM
remaining gets to a certain size/percentage. Increading RAM because
one feels there isn't enough space free therefore isn't really a
solution because Firefox will eventually expand to fill that new
space too. Although if it tries to leave a minimum percentage free,
the size of that minimum would increase.

But it would be great to see an explanation of how it really works
from one of the Firefox developers (one can dream).

> When I started using Linux, back in 1998, I did notice that Linux needed
> more memory than Windows to work right.

It did compared to mid-90s Windows. With XP and later the
difference switched the other way and it surprises me how much more
RAM Windows 10 can use doing nothing compared to ~100MB used after
modern Linux boots up with to a lightweight desktop. Older Linux
used much RAM less than that, but still more than a basic Win98
installation (the old PC I'm posting from with 80MB RAM dual-boots
to either).

> I had to improve my hardware
> because of that. My previous computer had 8 GB and it was not enough, it
> was swapping. So now I have a machine with 64.

I wouldn't know what to do with 8GB (well I know some people like
to run lots of VMs, but that's too much to think about in my
opinion).

>> I don't use web browsers to play video.
>
> No youtube? :-)

Only with youtube-dl (yes I know there's also the yt-dlp fork). To
find videos there I also avoid their Javascript-filled website and
use Invidious as well as "site:youtube.com" in DuckDuckGo.

--
__ __
#_ < |\| |< _#

On 2024-05-04 02:13, Computer Nerd Kev wrote:
> Carlos E.R. <robin_listas@es.invalid> wrote:
>> On 2024-05-03 01:23, Computer Nerd Kev wrote:
>>>
>>> The OP has noted now that the process that consumes their RAM is
>>> Chrome or Firefox. I've not seen a detailed description of why it
>>> happens, but I've long noted that Firefox seems to expand its RAM
>>> usage to the available space, but different from a memory leak in
>>> that it usually leaves a certain amount free. I assume that this
>>> in intended behaviour. I run current Firefox on a PC with 2GB RAM
>>> and I don't have it getting killed by the kernel, nor do I have
>>> problems with kernel crashes/reboots. I've also tried running
>>> recent Firefox on a PC with 512MB RAM and noticed that it performs
>>> much worse than with 2GB RAM, slowing down to a crawl while loading
>>> some websites, suggesting that it really does need more RAM in that
>>> case.
>>
>> Firefox uses a lot of memory.
>>
>> One problem is memory fragmentation. It reserves many chunks then
>> eventually frees many, but they may not being contiguous, so maybe they
>> can not be reused, so it requests more memory.
>
> It seems to me like it looks at the available RAM (not including
> swap space) and only works at freeing space when the free RAM
> remaining gets to a certain size/percentage. Increading RAM because
> one feels there isn't enough space free therefore isn't really a
> solution because Firefox will eventually expand to fill that new
> space too. Although if it tries to leave a minimum percentage free,
> the size of that minimum would increase.
>
> But it would be great to see an explanation of how it really works
> from one of the Firefox developers (one can dream).

If memory serves, my explanation above comes from a Chrome dev,
distorted through years of retelling it ;-)

....

--
Cheers, Carlos.

On 4 May 2024 10:13:26 +1000, Computer Nerd Kev wrote:

> ... as well as "site:youtube.com" in DuckDuckGo.

Hey, does that work? Because I think Google dropped support for it a while
back.

Lawrence D'Oliveiro <ldo@nz.invalid> wrote:
> On 4 May 2024 10:13:26 +1000, Computer Nerd Kev wrote:
>
>> ... as well as "site:youtube.com" in DuckDuckGo.
>
> Hey, does that work?

Sure, I'm using it more and more for websites where their own
search doesn't work without Javascript.

> Because I think Google dropped support for it a while back.

Searching for "Linux memory management site:youtube.com" works for
me right now with DDG and Google.

--
__ __
#_ < |\| |< _#

On 04/05/2024 01:13, Computer Nerd Kev wrote:
> I wouldn't know what to do with 8GB (well I know some people like
> to run lots of VMs, but that's too much to think about in my
> opinion).

Once you run a GUI and some modern python style code its amazing how
much ram crap apps can gobble up

I just upgraded from 8GB because the combination of 3GB of windows VM
plus a browser running a 3D printers webserver with embedded video, plus
a Python 3D slicer overwhelmed the machine and sent it into swap.

Firefox seems to spawn an infinite number of 'webkit' and 'isolated web
company' children

Especially with necessary ad blockers installed

Thunderbird is over half a GB

Orca slicer is 400GB+

Plus two instances of a linux game I am playing takes *hard* usage to
over 8GB/ I.e before cached files etc.

Now these are all things I want to do *at the same time*. Edit in
windows VM, pass 3D files to the slicer, monitor print progress via the
web interface, and play a linux game to while away the minutes waiting
for the 3D printer to turn out yet another 'close, but no cigar' print
that needs re-editing. Or while reading the news^H^H^H^H propaganda, or
Usenet, or getting email in.

All I can say is, thank Clapton Linux has a really good memory manager.

$ free -m
total used free shared buff/cache
available
Mem: 23909 7385 931 1044 15592
15090
Swap: 2047 0 2047

I can't control 3rd party apps, but I can control my machine's memory.

Up till buying a 3D printer 8GB was more than enough, but the constant
use of windows based 3D and 2D CAD programs plus the slicer and
firefoxes latest bloatery has taken it over the edge

--
For every complex problem there is an answer that is clear, simple, and
wrong.

H.L.Mencken

The Natural Philosopher <tnp@invalid.invalid> wrote at 05:47 this Friday (GMT):
> On 02/05/2024 14:50, James Harris wrote:
>> On 01/05/2024 11:53, The Natural Philosopher wrote:
>>> On 01/05/2024 11:32, James Harris wrote:
>>>> Not a question, just an observation.
>>>>
>>>> I say that Linux doesn't seem to handle memory well because my laptop
>>>> had 8GB RAM (which, frankly, Windows seems to find perfectly adequate
>>>> for a similar workload). Under Linux the RAM would fill up and then
>>>> swap space would be used. Then the machine would become largely
>>>> unresponsive - e.g. taking minutes to switch between windows.
>>>>
>>>> So I upgraded the RAM. It now has three times as much (i.e. 24GB)!
>>>> But even so, RAM has still steadily filled up until reaching the full
>>>> 24GB. What's more, it's now showing 4.8GB of swap space in use.
>>>>
>>>>
>>> That's not Linux. That's a rogue program with a memory leak...
>>
>> It may not be a memory leak but it's certainly a rogue program -
>> chrome/chromium in this case, though I've had similar and worse problems
>> with Firefox which seems to me more likely to have a genuine memory leak.
>>
>> However, the point of the original post was to vent a frustration about
>> the way that Linux handles such memory deletion when it happens.
>>
>>
> It is not hard to write a javascript program that eats all memory and
> place it on a website.
>
> I even encountered a website that silently ran some sort of proxy server
> in my browser and used up all my upload bandwidth.
>
>
> Linux at least protects you from such sites having root access by default...

I think sandboxing is /supposed/ to protect you from that.
--
user <candycane> is generated from /dev/urandom

The Natural Philosopher <tnp@invalid.invalid> wrote at 09:48 this Saturday (GMT):
> On 04/05/2024 01:13, Computer Nerd Kev wrote:
>> I wouldn't know what to do with 8GB (well I know some people like
>> to run lots of VMs, but that's too much to think about in my
>> opinion).
>
> Once you run a GUI and some modern python style code its amazing how
> much ram crap apps can gobble up
>
> I just upgraded from 8GB because the combination of 3GB of windows VM
> plus a browser running a 3D printers webserver with embedded video, plus
> a Python 3D slicer overwhelmed the machine and sent it into swap.
>
> Firefox seems to spawn an infinite number of 'webkit' and 'isolated web
> company' children
>
> Especially with necessary ad blockers installed
>
> Thunderbird is over half a GB
>
> Orca slicer is 400GB+
>
> Plus two instances of a linux game I am playing takes *hard* usage to
> over 8GB/ I.e before cached files etc.
>
> Now these are all things I want to do *at the same time*. Edit in
> windows VM, pass 3D files to the slicer, monitor print progress via the
> web interface, and play a linux game to while away the minutes waiting
> for the 3D printer to turn out yet another 'close, but no cigar' print
> that needs re-editing. Or while reading the news^H^H^H^H propaganda, or
> Usenet, or getting email in.
>
> All I can say is, thank Clapton Linux has a really good memory manager.
>
>
> $ free -m
> total used free shared buff/cache
> available
> Mem: 23909 7385 931 1044 15592
> 15090
> Swap: 2047 0 2047
>
> I can't control 3rd party apps, but I can control my machine's memory.
>
> Up till buying a 3D printer 8GB was more than enough, but the constant
> use of windows based 3D and 2D CAD programs plus the slicer and
> firefoxes latest bloatery has taken it over the edge

Yeah, but what other JS-compatible browsers exist? Chromium is also bad
with memory..
--
user <candycane> is generated from /dev/urandom

Computer Nerd Kev <not@telling.you.invalid> wrote at 00:13 this Saturday (GMT):
> Carlos E.R. <robin_listas@es.invalid> wrote:
>> On 2024-05-03 01:23, Computer Nerd Kev wrote:
[snip]
>>> I don't use web browsers to play video.
>>
>> No youtube? :-)
>
> Only with youtube-dl (yes I know there's also the yt-dlp fork). To
> find videos there I also avoid their Javascript-filled website and
> use Invidious as well as "site:youtube.com" in DuckDuckGo.

ytfzf is another good tool for YouTube downloading.
--
user <candycane> is generated from /dev/urandom

On 06/05/2024 16:30, candycanearter07 wrote:
> The Natural Philosopher <tnp@invalid.invalid> wrote at 09:48 this Saturday (GMT):
>> On 04/05/2024 01:13, Computer Nerd Kev wrote:
>>> I wouldn't know what to do with 8GB (well I know some people like
>>> to run lots of VMs, but that's too much to think about in my
>>> opinion).
>>
>> Once you run a GUI and some modern python style code its amazing how
>> much ram crap apps can gobble up
>>
>> I just upgraded from 8GB because the combination of 3GB of windows VM
>> plus a browser running a 3D printers webserver with embedded video, plus
>> a Python 3D slicer overwhelmed the machine and sent it into swap.
>>
>> Firefox seems to spawn an infinite number of 'webkit' and 'isolated web
>> company' children
>>
>> Especially with necessary ad blockers installed
>>
>> Thunderbird is over half a GB
>>
>> Orca slicer is 400GB+
>>
>> Plus two instances of a linux game I am playing takes *hard* usage to
>> over 8GB/ I.e before cached files etc.
>>
>> Now these are all things I want to do *at the same time*. Edit in
>> windows VM, pass 3D files to the slicer, monitor print progress via the
>> web interface, and play a linux game to while away the minutes waiting
>> for the 3D printer to turn out yet another 'close, but no cigar' print
>> that needs re-editing. Or while reading the news^H^H^H^H propaganda, or
>> Usenet, or getting email in.
>>
>> All I can say is, thank Clapton Linux has a really good memory manager.
>>
>>
>> $ free -m
>> total used free shared buff/cache
>> available
>> Mem: 23909 7385 931 1044 15592
>> 15090
>> Swap: 2047 0 2047
>>
>> I can't control 3rd party apps, but I can control my machine's memory.
>>
>> Up till buying a 3D printer 8GB was more than enough, but the constant
>> use of windows based 3D and 2D CAD programs plus the slicer and
>> firefoxes latest bloatery has taken it over the edge
>
>
> Yeah, but what other JS-compatible browsers exist? Chromium is also bad
> with memory..

They are all like that, as the manager said..

https://www.youtube.com/watch?v=k58knuvcQjQ

--
A lie can travel halfway around the world while the truth is putting on
its shoes.

On 5/1/24 3:05 PM, D wrote:
>
>
> On Wed, 1 May 2024, James Harris wrote:
>
>> Not a question, just an observation.
>>
>> I say that Linux doesn't seem to handle memory well because my laptop
>> had 8GB RAM (which, frankly, Windows seems to find perfectly adequate
>> for a similar workload). Under Linux the RAM would fill up and then
>> swap space would be used. Then the machine would become largely
>> unresponsive - e.g. taking minutes to switch between windows.
>>
>> So I upgraded the RAM. It now has three times as much (i.e. 24GB)! But
>> even so, RAM has still steadily filled up until reaching the full
>> 24GB. What's more, it's now showing 4.8GB of swap space in use.
>
> Sounds like some rouge application and not linux. I never had any ram
> problems with linux given 4, 8 and now 16 gb of ram and I run firefox
> and xfce. Even ran gnome on 4 gb ram and no problems.

Agreed ... Linux manages memory perfectly OK. His
prob is a rogue app, likely failing to de-allocate
some kind of (largish) buffers. He may need to
run valgrind, or at least sudo pmap -x <pid>
several times on suspected apps to see what's
going on. ps can kind-of be coerced into showing
the real usage, but it's a lot of flags. top can
also give you a fair picture.

Dynamic memory allocation - albeit ancient - can
be a pain.