[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: file systems



In <4DC27008.2080105@hardwarefreak.com>, Stan Hoeppner wrote:
>On 5/2/2011 5:54 PM, Boyd Stephen Smith Jr. wrote:
>> I'm slightly surprised by the results.  It's possible it was slightly
>> weighted toward JFS because of the "%CPU" and "Ops/%CPU" metrics, which I
>> don't think matter too much.
>
>As I mentioned previously, the only relevant graph of each set is the
>one at the top of each page, either MB/s or IOPS.

Ops/sec is always the first graph.
CPU% is always the second graph.
MB/sec is always the third graph.
Mail server simulation has a second MB/sec graph.
Efficiency (Ops/CPU%) is always the last graph.

>The other graphs are
>useless for any kind of ranking, and in fact will likely produce
>misleading rankings.
>If your ranking program is looking at tall bars,
>it will make an inverse ranking error WRT %CPU data, where the lower
>bars are better.

I converted the graphs to ballots manually, so I took into account that short 
bars where better on that graph.

>%CPU isn't relevant anyway as any 4+ core server
>shipped in the past 3 years has a huge excess of CPU/mem bandwidth in
>relation to IO.

Agreed.

>If you rank strictly based on the first graph of each test result page
>you'll see why I use the word "trounces".

Based only on Ops/sec (first graph), the results are:

1. jfs
2. ext4-nobarrier
3. xfs-nobarrier
4. ext3
5. ext4
5. xfs
(tie)
7. btrfs-nocow
8. ext3-barrier
9. btrfs

Ballots used attached.
-- 
Boyd Stephen Smith Jr.                   ,= ,-_-. =.
bss@iguanasuicide.net                   ((_/)o o(\_))
ICQ: 514984 YM/AIM: DaTwinkDaddy         `-'(. .)`-'
http://iguanasuicide.net/                    \_/
9 1
1 3 4 8 9 7 6 5 1 2 0
1 9 4 7 8 6 3 5 1 2 0
1 9 4 8 3 6 2 1 7 5 0
1 7 4 1 9 3 6 5 2 8 0
1 7 4 9 1 6 2 5 3 8 0
1 9 7 4 1 6 5 2 8 3 0
1 3 7 4 1 6 5 8 9 2 0
1 3 4 6 1 2 5 7 8 9 0
1 1 4 3 2 6 5 7 8 9 0
1 7 4 8 9 1 3 2 6 5 0
1 9 8 7 4 3 1 2 6 5 0
1 1 4 7 9 6 8 3 2 5 0
1 8 7 9 4 1 3 2 6 5 0
1 9 7 4 1 8 3 2 6 5 0
1 7 9 8 4 1 2 3 5 6 0
1 6 5 9 1 8 7 4 3 2 0
1 6 3 7 4 8 1 9 2 5 0
1 4 3 9 7 1 6 8 2 5 0
0
"ext3"
"ext3-barrier"
"ext4"
"ext4-nobarrier"
"btrfs"
"btrfs-nocow"
"jfs"
"xfs"
"xfs-nobarrier"
"Linux File Systems"

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: