• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Try Btrfs if you are installing to an ssd

johnspack

Here For Good!
Joined
Oct 6, 2007
Messages
6,072 (0.94/day)
Location
Nelson B.C. Canada
System Name System2 Blacknet , System1 Blacknet2
Processor System2 Threadripper 1920x, System1 2699 v3
Motherboard System2 Asrock Fatality x399 Professional Gaming, System1 Asus X99-A
Cooling System2 Noctua NH-U14 TR4-SP3 Dual 140mm fans, System1 AIO
Memory System2 64GBS DDR4 3000, System1 32gbs DDR4 2400
Video Card(s) System2 GTX 980Ti System1 GTX 970
Storage System2 4x SSDs + NVme= 2.250TB 2xStorage Drives=8TB System1 3x SSDs=2TB
Display(s) 1x27" 1440 display 1x 24" 1080 display
Case System2 Some Nzxt case with soundproofing...
Audio Device(s) Asus Xonar U7 MKII
Power Supply System2 EVGA 750 Watt, System1 XFX XTR 750 Watt
Mouse Logitech G900 Chaos Spectrum
Keyboard Ducky
Software Archlinux, Manjaro, Win11 Ent 24h2
Benchmark Scores It's linux baby!
I did multiple tests in vms, and found btrfs to be significantly faster than ext4 on an ssd. Not recommended for storage drives.
Ubuntu runs stupid fast on it, and if you have 32gbs ram or more, you don't need the swap file. Even 16 would probably do it.
 
This isn't new known for 3 years
 
I did multiple tests in vms, and found btrfs to be significantly faster than ext4 on an ssd. Not recommended for storage drives.
Ubuntu runs stupid fast on it, and if you have 32gbs ram or more, you don't need the swap file. Even 16 would probably do it.
Do you have numbers to back that up? I would be intrigued if you do. Phoronix has short article on the matter and it looks like the only case where it was faster was using SQLite, it was as fast or slower with all the other benchmarks. I would expect any machine with more memory to go faster regardless of the type of file system because the file system and certain files are more likely to be cached.

https://www.phoronix.com/scan.php?page=article&item=linux-412-fs&num=2
 
It is indeed good, but the problem is the repair software when somerhing goes wrong, it's not as good as in EXT4.
 
Phoronix results don't agree. They showed ext4 to be the most well rounded. Also, there is a steady stream of dead btrfs partitions that pass through the opensuse forum. Just changing the fs isn't going to transform your distro into "stupid fast". And why would changing the fs have ANY bearing on whether you need a swap file for your PC or not? That's application dependent.
 
And why would changing the fs have ANY bearing on whether you need a swap file for your PC or not?
Forget the part where swap gets its own partition on a disk and isn't even stored on a file system on *nix machines.
That's application dependent.
It's memory usage dependent in the *nix world. They're just like any other memory page which is probably why being its own partition makes sense. It's easy to map memory addresses to addresses on a physical disk while maintaining relatively reasonable performance. No file system overhead for virtual memory.
 
Alright, not new, but much more stable now. I've only been pushing linux for like six months now so, cool it. Before installing ubuntu on my main rig, I tested 2 identical installs of kubuntu in vms, running off the same ssd. Only one at a time. Results were repeatable at least 20x. btrfs-tools do exist now, no reason not to install. Tests, first is ext4, 2nd is btrfs:
kubuntuext4.png

btrfsbench.png
 
That looks like the VM host doing caching like a pro. That doesn't look like anything to do with btrfs. Do the same benchmarks on a machine outside of a VM without write-back caching the size of system memory and we'll have a good discussion but, this just looks like vmware getting creative. If anything, those constantly low points probably show the actual file system performance without caching. Even my RAID-0 with SSDs maintains closer to what you would expect from two SATA-3 devices in RAID-0. No file system will mysteriously make a drive faster than it actually is.
1519694975843.png
 
I'll try some raw numbers and post those. I'll have to install another kubuntu on my 2nd equal ssd I guess. I don't understand why you can't see the difference, but okay.
Still compiling a list of windows games that run just fine under winehq wine right now... much more important....Oh, here's my ancient old system using btrfs raw:
raw1.png
 
Linux has always mapped swap to a dedicated partition ... you have to work extra hard to put in on a loopback device on a mounted fs, which is suicide in low mem conditions ...
 
I also have Phoronix suites installed. In the case of btrfs though, there is no swap file. Not using tmp dir... I've never seen numbers like this under windows. Except with magician. Just under 600mbs isn't cached though.
Raw numbers. Well I can't argue anymore, I'm just going to tweak this bi0tch, and take over the world....
 
I suppose it could be faster within a VM for some unexpected reason / interaction with the particular host system.
 
Yes, it's obvious vmware is doing some caching work... I'll have to do a clean install raw using ext4 to compare fully. I'm still happy with btrfs performance on my raw install, no issues, no errors. And I'm pushing my install. I want to break it, to see it's limits. So far btrfs and ubuntu just friggin rock.....
 
No file system will mysteriously make a drive faster than it actually is.

btrfs wiki page says that it does actually increase the throughput if that helps.

"SSD (flash storage) awareness (TRIM/Discard for reporting free blocks for reuse) and optimizations (e.g. avoiding unnecessary seek optimizations, sending writes in clusters, even if they are from unrelated files. This results in larger write operations and faster write throughput)"

https://btrfs.wiki.kernel.org/index.php/Main_Page
 
Yea even on My HDD it boots faster and access faster. Don't know why Linux doesn't have it as default
 
According This This Benchmark:

https://www.phoronix.com/scan.php?page=article&item=linux414-fs-compare&num=2

View attachment 101577View attachment 101578

It looks like Where btrfs Performs Well Is in The Sequential Read Which would make sense why it is booting and accessing things faster. where it runs slower is the sequential writing to the disk, it performs other functions at this time, which would also make sense why it would be slower.

At that point it wouldn't be worth using it considering the reads aint drastically better and the writes are drastically worse than the other formats.
 
As I mentioned before btrfs just needs some work. It's still new to a lot of people. I used it in mint 17.3 and it had some issues with reading and writing to the drive
 
I have somewhat felt this at least on my laptop with suse/centos/fedora in btrfs and mint/ubuntu/arch/parrot/gentoo in ext4. The boot time for fedora is faster than mint, however writing is much faster in mint.
 
That's because MINT is better optimized... :pimp:
 
Back
Top