moonpiedumplings, moonpiedumplings@programming.dev

Instance: programming.dev
Joined: 2 years ago
Posts: 14
Comments: 187

RSS feed

Posts and Comments by moonpiedumplings, moonpiedumplings@programming.dev

Unless you are running at really large scales, or really small scales and trying to fit stuff that quite fit, memory compression may not be significant enough of an optimization to spend a lot of time experimenting a lot. But I’m bored and currently on an 8 GB device so here are my thoughts dumped out from my recent testing:

Zram vs Zswap (can be done at hypervisor or at host):

  • One or the other is commonly enabled on many modern distros. It is a perfectly reasonable position to simply use the distro’s defaults and not push it any further
  • Zram has much, much better compression, but suffers from LRU inversion. Essentially after zswap is full, fresh pages (memory) goes to the swap instead. Since these pages will probably be needed, it will be slower to get them from the disk then to get them from zram.
  • Zswap has much, much worse compression but cold, unused pages are moved to swap automatically, freeing up space
  • I am investigating ways to get around the above. See my thoughts on this and other differences here: https://github.com/moonpiedumplings/moonpiedumplings.github.io/blob/main/playground/asahi-setup/index.md#memory-optimization

Kernel same page merging (KSM) (would be done at hypervisor level) (esxi also has an equivalent feature called something different):

  • Only really efficient if you have lots of the same virtual machines
  • Used to overcommit (promise more ram than you physically have)
    • Dangerous, but highly cost saving. Many cheap VPS providers do this in order to save money. You can run four 8 GB vps on 24 GB of ram and take a semi-safe bet that not all of the memory will be used.

In my opinion, the best thing is to enable zram or zswap at the virtual machine level and kernel same page merging at the hypervisor level, assuming you take into account and accept the marginal security risk and slightly weaker isolation that comes with KSM. There isn’t any point running zswap at two layers, because the hypervisor is just gonna spend a lot of time trying to see if it can compress stuff that’s already been compressed. Than KSM deduplicates memory across hosts. Although you may actually see worse savings overall if zram/zswap compression is only semi-deterministic and makes deduplication ahrder.

I agree with the other commenter as well about zram being weird with some workloads. Like I’ve heard of I think it was blender interacting weirdly with zram since zram is swap, making less total memory available in ram, whereas zswap compresses memory. If you really need to know you gotta test.


That’s not quite true. Virtualbox is free but the extension pack is not. It says on the website that it’s under a different license.

Just don’t get it from the website but from a distro’s repos instead and you’ll be fine. Distros usually patch out telemetry as well.

But yeah, Oracle and similar schemes are why software installation is so restricted on corporate devices. It’s basically ransomware, freeware that people are willing to sue over.

Edit: it should be noted that charging people for licensed software in a corporate environment is okay. I have heard stories of Oracle making people buy licences for EVERY computer even if only one person downloaded the software…


Does the script attempt to run though? If linkedin runs this and other scripts it would explain why the site is so bloated.


Does this work on firefox? Does ublock origin block this?

Is this why linkedin eats so damn much ram. It eays 300 mb for a single tab. I opened 3 linkedin tabs and it lagged my entire computer.


Sometimes copyrighted stuff gets dmca’d?



Rust Coreutils Continues Working Toward 100% GNU Compatibility, Proving Trolls Wrong

98 comments

Phoronix, you are the trolls.



Have you considered that the reason why your mail server is trusted is because it’s been around for 20 years?

Have you tried to set up mail from scratch on a new domain/ip recently?


these ones: https://www.etsy.com/shop/SoHexy ?

I think I’m in love. They have such great variety, and the artstyle is so neat. And I love stickers because they are such great conversation starters.


Yes, there are often extensions to static site generators that have this. I like quarto because this is built into the static site generator itself and is just a toggle option but it is a common feature.


In the old days, university IT put essentially no access controls on their networks, so students’ dorm computers were completely exposed to the internet

Dorm ethernet works this way for me right now. It’s how I host some stuff. I only get 100 mb/s per port though. I’ve bonded two ports to get 200 total.


Because the extensions replaced wordpress’ sitebuilder/editor. If I were to get rid of the extensions I would basically have to recreate the site anyways so I might as well switch away from wordpress.


It powers lichess.org, who have made multiple blogposts about how happy they are with it.

Lichess is a FOSS chess server that somehow manages to compete with chess.com proprietary, distributed, milticloud kubernetes setup from a single VPS. According to them, scala helps.




mp3 is still the best in terms of compatibility. Basically anything can play it.

m4a is better than mp3 every eay and fills the same usecases. For the same size as an mp3, an m4a can offer you better quality. For a smaller size, m4a can offer you the same quality.


It’s possible that the newer appimage depends on an external newer version of the glibc library, whereas the older appimage bundles it into the appimage.

I tracked down the relevant github issue. That issue is closed and it looks like they implemented a fix already.

Try 4.0.6


RSS feed

Posts by moonpiedumplings, moonpiedumplings@programming.dev

Comments by moonpiedumplings, moonpiedumplings@programming.dev

Unless you are running at really large scales, or really small scales and trying to fit stuff that quite fit, memory compression may not be significant enough of an optimization to spend a lot of time experimenting a lot. But I’m bored and currently on an 8 GB device so here are my thoughts dumped out from my recent testing:

Zram vs Zswap (can be done at hypervisor or at host):

  • One or the other is commonly enabled on many modern distros. It is a perfectly reasonable position to simply use the distro’s defaults and not push it any further
  • Zram has much, much better compression, but suffers from LRU inversion. Essentially after zswap is full, fresh pages (memory) goes to the swap instead. Since these pages will probably be needed, it will be slower to get them from the disk then to get them from zram.
  • Zswap has much, much worse compression but cold, unused pages are moved to swap automatically, freeing up space
  • I am investigating ways to get around the above. See my thoughts on this and other differences here: https://github.com/moonpiedumplings/moonpiedumplings.github.io/blob/main/playground/asahi-setup/index.md#memory-optimization

Kernel same page merging (KSM) (would be done at hypervisor level) (esxi also has an equivalent feature called something different):

  • Only really efficient if you have lots of the same virtual machines
  • Used to overcommit (promise more ram than you physically have)
    • Dangerous, but highly cost saving. Many cheap VPS providers do this in order to save money. You can run four 8 GB vps on 24 GB of ram and take a semi-safe bet that not all of the memory will be used.

In my opinion, the best thing is to enable zram or zswap at the virtual machine level and kernel same page merging at the hypervisor level, assuming you take into account and accept the marginal security risk and slightly weaker isolation that comes with KSM. There isn’t any point running zswap at two layers, because the hypervisor is just gonna spend a lot of time trying to see if it can compress stuff that’s already been compressed. Than KSM deduplicates memory across hosts. Although you may actually see worse savings overall if zram/zswap compression is only semi-deterministic and makes deduplication ahrder.

I agree with the other commenter as well about zram being weird with some workloads. Like I’ve heard of I think it was blender interacting weirdly with zram since zram is swap, making less total memory available in ram, whereas zswap compresses memory. If you really need to know you gotta test.


That’s not quite true. Virtualbox is free but the extension pack is not. It says on the website that it’s under a different license.

Just don’t get it from the website but from a distro’s repos instead and you’ll be fine. Distros usually patch out telemetry as well.

But yeah, Oracle and similar schemes are why software installation is so restricted on corporate devices. It’s basically ransomware, freeware that people are willing to sue over.

Edit: it should be noted that charging people for licensed software in a corporate environment is okay. I have heard stories of Oracle making people buy licences for EVERY computer even if only one person downloaded the software…


Does the script attempt to run though? If linkedin runs this and other scripts it would explain why the site is so bloated.


Does this work on firefox? Does ublock origin block this?

Is this why linkedin eats so damn much ram. It eays 300 mb for a single tab. I opened 3 linkedin tabs and it lagged my entire computer.


Sometimes copyrighted stuff gets dmca’d?



Rust Coreutils Continues Working Toward 100% GNU Compatibility, Proving Trolls Wrong

98 comments

Phoronix, you are the trolls.



Have you considered that the reason why your mail server is trusted is because it’s been around for 20 years?

Have you tried to set up mail from scratch on a new domain/ip recently?


these ones: https://www.etsy.com/shop/SoHexy ?

I think I’m in love. They have such great variety, and the artstyle is so neat. And I love stickers because they are such great conversation starters.


Yes, there are often extensions to static site generators that have this. I like quarto because this is built into the static site generator itself and is just a toggle option but it is a common feature.


In the old days, university IT put essentially no access controls on their networks, so students’ dorm computers were completely exposed to the internet

Dorm ethernet works this way for me right now. It’s how I host some stuff. I only get 100 mb/s per port though. I’ve bonded two ports to get 200 total.


Because the extensions replaced wordpress’ sitebuilder/editor. If I were to get rid of the extensions I would basically have to recreate the site anyways so I might as well switch away from wordpress.


It powers lichess.org, who have made multiple blogposts about how happy they are with it.

Lichess is a FOSS chess server that somehow manages to compete with chess.com proprietary, distributed, milticloud kubernetes setup from a single VPS. According to them, scala helps.




mp3 is still the best in terms of compatibility. Basically anything can play it.

m4a is better than mp3 every eay and fills the same usecases. For the same size as an mp3, an m4a can offer you better quality. For a smaller size, m4a can offer you the same quality.


It’s possible that the newer appimage depends on an external newer version of the glibc library, whereas the older appimage bundles it into the appimage.

I tracked down the relevant github issue. That issue is closed and it looks like they implemented a fix already.

Try 4.0.6


It’s mostly a self hosted thing I think.

Also, it’s more of a “meta service”, as it’s a package thay contains many relevant other services, like opencloud (nextcloud/gdrive alternative), element/matrix and more. It’s designed to be the whole package.

You can also consider deploying or paying for SAAS versioms of each software individually, which may be better because, as you’ve noticed, the docs for some of the German software is all in German.

Also, it does seem to have a SAAS:

Our SaaS offering is available to organisations in Germany with a minimum of 500 users, but on-premise installations have no minimum user requirements.

From the faq.


This should just work without any alterations though.

But don’t disable DHCP for wan, that’s probably not it. That’s needed for connectivity