NixOS Takes Action After 1.2GB/s ZFS Encryption Speed Drops To 200MB/s With Linux 5.0+

[starshipeleven]

Please, for the lord of god, stop feeding him/her(it?).

Ignore us both then.

[skeevy420]

No sane enthusiast gamer would use flippin' Linux. "Sunday gamers" are not it. Problem is not only relative lack of games, problem is inferior performance of existing games (compared to windows) and relative lack of 3rd party software important to gamers. Windows platform has plethora of fine-tuning software for over-clocking for example.
Show me Linux-equivalent of MSI Afterburner for example? Or Linux-specific support software for top-of-the-line gaming peripherals? Keyboards, mouse, headsets, joysticks, pedals, throttles..

Lets top it off with the "issue" where Linux-ported games usually lack anti-cheat engines (Valve's games being sole exception) and are lagging multiple versions behind Windows variants.

It really depends on the kinds of games you like to play if Linux will work for you or not. I'm currently playing Hitman 2 and it's working great. PS4 controller works with it and other Wine games OOTB with the kernel driver (prompts are for Xbox controller sometimes...common PC problem and isn't Linux exclusive). All the games that I like to play & own, slower single player games, work just fine. YMMV if you play more online multiplayer games.

There's Wattman-gtk and a few other programs like that. I don't think we have anything that provides what Afterburner does in a single package, though, and it's been a while since I've used it so more features might have been added, but IIRC, all of it's functionality can be found using multiple packages.

Hardware support on Linux has never been the greatest, but it is getting better and there are 3rd party projects that do bring some support to various gaming hardware if one really searches for them...like libratbag/piper, that xneo driver for xbox one controllers (can't remember it's full name), the r600g header, sc-controller...

[lucrus]

starshipeleven we can surely ignore both of you, but while what you say sounds generally reasonable and many of us agree with you, so ignoring you is undesirable, what the troll says is blatantly a pile of dogshit to everyone even if you don't debunk that crap: we already know he/she/it is a troll.

[useless]

starshipeleven we can surely ignore both of you, but while what you say sounds generally reasonable and many of us agree with you, so ignoring you is undesirable, what the troll says is blatantly a pile of dogshit to everyone even if you don't debunk that crap: we already know he/she/it is a troll.

Thank you. I've encountered some of starshipeleven past posts kind of insightful (not in this thread, particularly), so I don't wan't to just ignore him. Replying to a troll only encourages it to troll us even more.

[aht0]

Sorry but console controllers are not exactly enthusiast gaming. It's "couch, TV and a beer couple of hours over a weekend" sort of thing. You are out of luck with programmable gaming peripherals, most PCI Express video capture cards (at least OBS Studio is open software) and have major head ache in the form of driver breakage following updates.

Afterburner does lots of things at the same time. Starting from regulating GPU fans according to temperature, ending with a FPS/CPU/MEM overlays on your screen. Add in everything in between like recording videos, taking screenshots, overclocking GPU and so forth. In order to replace it, you'd need about dozen different utilities.
Now, tell me the probability that you'd run into some sort of issues with just one (if not more) of the replacement utilities? I'd say it's close to infinity.

Imagine you set painstakingly everything up. You have about 15 different utilities properly configured. Even manage to game some days. Then system update happens - and you'd spend next 2x24h trying to figure bunch of issues - starting from f-ed up graphics driver to some segfaulting gaming utility. Joy, isn't it? 2 such incidents later, you'll give up and install windows.

Enough of OT now

[scineram]

No one gives two fucks about Oracle.

[Darksurf]

Honestly, I think Redhat's new Filesystems will probably kill ZFS. Redhat has a better reputation than Oracle and corporate users will not want to store important data on some makeshift Linux bridge.

Have you looked into Redhats "New FS". It's not even a filesystem.. it's a terrible "makeshift Linux bridge" using XFS and a daemon. It wouldn't be wise to replace ZFS with something like that. I'm more willing to place my chips in BTRFS unless redhat rethinks that atrocity.

[birdie]

Have you looked into Redhats "New FS". It's not even a filesystem.. it's a terrible "makeshift Linux bridge" using XFS and a daemon. It wouldn't be wise to replace ZFS with something like that. I'm more willing to place my chips in BTRFS unless redhat rethinks that atrocity.

There's a much better FS which has been brewing for the past several years: https://www.patreon.com/bcachefs

It can replace all of ext4, ZFS, Btrfs, XFS and JFS.

[oiaohm]

systemd does not run with the 3.14 android kernel because of the unimplemented syscall 279,

https://github.com/systemd/systemd/blob/master/README

This readme contains the list of required kernel parts. Please get your facts right.

ASOP 3.14 the open source android kernel from google has all the features on by default to run systemd. This is case vendor has turned something off. Interesting enough this add complexity to rolling out android updates as well as running systemd.

279 is memfd_create. Of course you arguement gets more into fail once you realise one of the Google Android Kernel developers made that syscall and had it added to the Linux kernel. Yes memfd_create appears in android kernels before it appears in mainline.

So what ever device has that issue has a broken android kernel not built to android specifications. Do you fell lucky punk with android updates basically.

[oiaohm]

Have you looked into Redhats "New FS". It's not even a filesystem.. it's a terrible "makeshift Linux bridge" using XFS and a daemon. It wouldn't be wise to replace ZFS with something like that. I'm more willing to place my chips in BTRFS unless redhat rethinks that atrocity.

Not quite as makeshift as it first appears. History of XFS at the start it had XLV

https://techpubs.jurassic.nl/manuals/0530/admin/XFS_AG/sgi_html/ch01.html

Yes when you find old manuals of XFS it talks about XLV yes this is a Logical volume manager designed to work with XFS. When XFS came to Linux it lost its logical volume management XLV.

https://lwn.net/Articles/747633/ this new tricks work on XFS is restoring functionality that did not make it across the first time as well as adding stacking. Yes XFS on IRIX where it came from was able to see though the XLV under it where ported version to Linux could not.

https://stratis-storage.github.io/

Really in a lot of ways is a front end and is really updating the LVM system in Linux.

ZFS and BTRFS both have integrated volume management. XFS model is different where file system and logical volume manager are meant to be two different things with a proper communication framework in the middle of course that has been missing on Linux.

Yes I really do hope XFS target to be able to directly mount file system image without needing loopback will come reality for more than xfs.

Its funny how much that stratis does with xfs is just replacing functionality that was not ported to Linux in the first place. If xfs had been fully implemented with logical volume management support on Linux btrfs and zfs would have been lower interest.