If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
If the ext3 configuration used in the test was indeed not equivalent to the configuration of ext4 and btrfs then the results are _really_ interesting for btrfs. If its performance will not regress, it will beat the other filesystems hands on, based on the functionality. It doesnt look like ext4 will live long.
On a separate note, does anyone know what the advantages of using zlib compression on btrfs are?
Imagine you have an external array. The link to the array is about ~120MB/s (1 gigabit). The data you write to the array is not compressed. Your maximum write speed will be something less than the link speed.
zlib compression allows you to write more across a link per second. Its actually a speed optimization. Now you are writing 120MB/s of compressed data across the link.
Assuming the cpu load is negligible in your case, zlib compression can help with bandwidth throttled workloads.
This is the driver behind compression under ZFS as well.
I agree with the others. SSD's are not universally affordable and most people are still looking at performance with 7200rpm 1TB+ or bigger drives.
I would really like a retest with a standard 7200, and maybe even a laptop 5400 drive. (As laptops outsell desktops now).
The existing filesystems are designed to work with a rotating HD anyways. Their algorithms take into account latency and head movement, all of which is non-existent on SSDs.
Also, while you are at it, please take a look at JFS, and Reiser4. Get the zen-kernel (also available as a ppa) which supports reiser4. With Reiser4 possibly going into mainline I think it would be fair to compare it.
Finally, test reiser4 with lzo and gzip compression vs btrfs w/zlib.
Thanks for the work. I don't want to sound ungrateful. But these tests with SSD's don't make much sense, when SSD filesystems are designed differently.
Again. Strange comparison based upon Ubuntu system. Why?
It should be noticed, that Ubuntu Ext3 does not use barriers by default in order to look much faster. But this is big lie, putting users data into the danger!
Typical Ext3 speed on distribution, that care safety of users data, would be much slower in these graphs.
Totally i agree.
Under my point of view the policy of phoronix of benchmarking always at default settings creates unfair conditions (and misinformation to non aware lectors). On this case 1 filesystem is running without barriers whereas the other have barriers enabled.
Agreed, maybe Phoronix should choose better Linux distribution to benchmark file systems performance with default settings.
Otherwise those graphs are not file systems benchmarks but rather Ubuntu experience benchmarks -> completely worthless for broader audience.
Why not rename Phoronix test suite to Ubuntu test suite?
Because it's not only made for Ubuntu and it run on other distributions and systems as well. However, I'd like to see generic kernel benchmarks with the same mount options if possible.
Sure, but what kernel version are you using? Phoronix have already pointed out very recently that there is a serious ext4 performance regression in the latest kernel (2.6.32) compared to previous releases.
This regression is just a change in Ext4.
Originally Posted by jbrown96 View Post
It's disappointing to see how Ext4 seems to be going backwards. I know that the mount options make a huge difference with ext4, but I haven't really been able to follow them very well. My computer is a laptop, so it has a UPS (the battery) built in. Does anyone know some mount options that would improve performance. I'm not worried about data integrity since I have a battery.
You didn't read previous comments, did you? If Ext3 had different mount options then Ext4 what's the point in such comparison and then you base on what saying Ext4 seems to be going backwards?
Judging from this particular "benchmark" it seems whole file systems development that happened during all those days (after EXT3 was released) was purely pointless - is this what we're supposed to read as conclusion? (you know - the point of making benchmarks is to deliver conclusion other than 'benchmark is inappropriate').
Michael, please try to be more vanilla.
Or the opposite - tweak configurations so that meaningful conclusions can be drawn from it.
As far as I know, SSD mode is activated by default nowadays with btrfs. Are you sure it was disabled?
Anyways, here's my future predictions:
- btrfs will not be avaible with Ubuntu 10.04
- btrfs will be avaible as an option for Ubuntu 10.10 which ships Linux 2.6.35/2.6.36
- I will use btrfs with Ubuntu 10.10 and it will seriously rock
- btrfs will become the default filesystem and replace ext4 in most mainstream distributions in early 2011
Thanks. Want to share when the messiah will be down to greet us? It's a filesystem benchmark, what's that got to do with when ubuntu users will be blessed with btrfs? I bet you're one of these people that upgraded to Ubuntu xx.xx without even knowing what the difference was.
I'm tired of these tests where default options are used everywhere. What's the point? Show us the potential of these filesystems not just the fact that no configuration = crap performance.