New Benchmarks Of OpenSolaris, BSD & Linux

[phoronix]

Phoronix: New Benchmarks Of OpenSolaris, BSD & Linux

Earlier today we put out benchmarks of ZFS on Linux via a native kernel module that will be made publicly available to bring this Sun/Oracle file-system over to more Linux users. Now though as a bonus we happen to have new benchmarks of the latest OpenSolaris-based distributions, including OpenSolaris, OpenIndiana, and Augustiner-Schweinshaxe, compared to PC-BSD, Fedora, and Ubuntu.

http://www.phoronix.com/vr.php?view=15476

[joffe]

What on earth is 'Augustiner-Schweinshaxe'? Any relation to SchilliX? Completely new?

[Michael]

What on earth is 'Augustiner-Schweinshaxe'? Any relation to SchilliX? Completely new?

Yeah, it's codename for new OS derived from OpenSolaris code-base Also just happens to be one of the best names too mmmm Schweinshaxe und bier!

[curaga]

Pic missing on the first page? Also, mute -> moot on the last page.

[Lennie]

You are doing single-disk performance and you say: why use ZFS ?

Maybe because it is faster on many devices ?:

BTRFS on Ubuntu versus ZFS on FreeBSD:

             ZFS              BtrFS
1 SSD      256 MiByte/s     256 MiByte/s
2 SSDs     505 MiByte/s     504 MiByte/s
3 SSDs     736 MiByte/s     756 MiByte/s
4 SSDs     952 MiByte/s     916 MiByte/s
5 SSDs    1226 MiByte/s     986 MiByte/s
6 SSDs    1450 MiByte/s     978 MiByte/s
8 SSDs    1653 MiByte/s     932 MiByte/s
16 SSDs   2750 MiByte/s     919 MiByte/s

http://marc.info/?l=linux-btrfs&m=128101763830740&w=2

[Lennie]

Maybe I should add something to the previous post:

after some tuning it got much better:

Reference figures:
16* single disk (theoretical limit): 4092 MiByte/s
fio data layer tests (achievable limit): 3250 MiByte/s
ZFS performance: 2505 MiByte/s

BtrFS figures:
IOzone on 2.6.32: 919 MiByte/s
fio btrfs tests on 2.6.35: 1460 MiByte/s
IOzone on 2.6.35 with crc32c: 1250 MiByte/s
IOzone on 2.6.35 with crc32c_intel: 1629 MiByte/s
IOzone on 2.6.35, using -o nodatasum: 1955 MiByte/s

But still not as fast, nodatasum is not something you do I production I assume.

[kebabbert]

I would like to see benchmarks on LARGE servers. Solaris has always been targeting large servers with hundreds of CPUs and many many drives and much RAM. Linux has always been developed on desktop PCs and targeting 2-4 cpu servers.

If there where benchmarks on large servers with many drives and many cpus, we would see that Linux is far behind Solaris.

The above post only confirms what I say. But that is only a measly 16 disk drives involved. I would want to see benchmarks with 48 drives, or 96 drives or more.

[kraftman]

I would like to see benchmarks on LARGE servers. Solaris has always been targeting large servers with hundreds of CPUs and many many drives and much RAM. Linux has always been developed on desktop PCs and targeting 2-4 cpu servers.

If there where benchmarks on large servers with many drives and many cpus, we would see that Linux is far behind Solaris.

If we would see HPC benchmarks then Solaris could go home then. Wait a minute, it already did!

The above post only confirms what I say. But that is only a measly 16 disk drives involved. I would want to see benchmarks with 48 drives, or 96 drives or more.

It shows zfs was faster then btrfs in this configuration, nothing more.

[Lennie]

Totally agree about the large clusters. They pretty much all run Linux or are specialized in doing HPC:

http://www.top500.org/stats/list/36/osfam

[kebabbert]

If we would see HPC benchmarks then Solaris could go home then. Wait a minute, it already did!

Great, show us HPC benchmark comparisons between Solaris and Linux you talk about. If you do not post them here, you are a FUDer.

It shows zfs was faster then btrfs in this configuration, nothing more.

Oh yes? Than I can say the same:

"The HPC benchmarks only shows Linux was faster than Solaris in that configuration. Nothing more."

But I doubt those benchmarks exists. You have confessed you FUD in one post, so this is just probably more FUD from you.