Why is Reiser4 slower then ReiserFS v3

Why is Reiser4 slower then ReiserFS v3

[Dark Shadow]

I am running a Linux from Scratch system and I just upgraded to Linux
kernel 2.6.10-cko1 (first time away from a vanilla kernel) in order to
try out Reiser4 but after formatting a drive I noticed it is slower
here is the best test I could run

I have three hard drives so I took a file from one and copied it to
the others and timed it
source drive /dev/hda Reiser3 Western Digital 40gb 7200rpm
target1 drive /dev/hdb Reiser4 Western Digital 80gb 7200rpm
target2 drive /dev/sda Reiser3 Seagate 160gb 7200rpm (SATA but still
same rpm as rest so it should be the same)



time cp ~/800mb.file /target1
real    0m41.409s
user    0m0.010s
sys     0m4.364s


time cp ~/800mb.file /target2
real    0m38.318s
user    0m0.017s
sys     0m5.627s


What formatting command would I use the get best speed for the
following situation
capturing raw uncompressed 720x480x29.97 video from my pci tv card.
The best I got with reiserfs v3 was 640x480x29.97 and it would end up
making a 50gb file in 64 minutes.

Re: Why is Reiser4 slower then ReiserFS v3

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 1137 bytes --]

On Mon, 27 Dec 2004 13:38:12 MST, Dark Shadow said:

> I have three hard drives so I took a file from one and copied it to
> the others and timed it
> source drive /dev/hda Reiser3 Western Digital 40gb 7200rpm
> target1 drive /dev/hdb Reiser4 Western Digital 80gb 7200rpm
> target2 drive /dev/sda Reiser3 Seagate 160gb 7200rpm (SATA but still
> same rpm as rest so it should be the same)

You may wish to run 'hdparm -T -t' on each drive and see what the *raw* speed
is.  All drives are not created equal... ;)

> time cp ~/800mb.file /target1
> real    0m41.409s
> user    0m0.010s
> sys     0m4.364s

> time cp ~/800mb.file /target2
> real    0m38.318s
> user    0m0.017s
> sys     0m5.627s

Similarly, you should try each one 3-5 times and get an average (for
starters, if you have more than 800M of memory, the second time around it
may all still be in cache, so the second time gets a hot-cache boost). It
may be useful to run the command once and *ignore* its times, and then
re-run the command 3 times and average those results (so all 3 times you
actually *use* start from the same "previous command just finished" cache state).

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Why is Reiser4 slower then ReiserFS v3

[Hans Reiser]

Valdis.Kletnieks@vt.edu wrote:

>On Mon, 27 Dec 2004 13:38:12 MST, Dark Shadow said:
>
>  
>
>>I have three hard drives so I took a file from one and copied it to
>>the others and timed it
>>source drive /dev/hda Reiser3 Western Digital 40gb 7200rpm
>>target1 drive /dev/hdb Reiser4 Western Digital 80gb 7200rpm
>>target2 drive /dev/sda Reiser3 Seagate 160gb 7200rpm (SATA but still
>>same rpm as rest so it should be the same)
>>    
>>
>
>You may wish to run 'hdparm -T -t' on each drive and see what the *raw* speed
>is.  All drives are not created equal... ;)
>  
>
This is correct.  You should see negligible difference between the two 
for large files on the same drive.  Deletions however, see next email....

>>time cp ~/800mb.file /target1
>>real    0m41.409s
>>user    0m0.010s
>>sys     0m4.364s
>>    
>>
>
>  
>
>>time cp ~/800mb.file /target2
>>real    0m38.318s
>>user    0m0.017s
>>sys     0m5.627s
>>    
>>
>
>Similarly, you should try each one 3-5 times and get an average (for
>starters, if you have more than 800M of memory, the second time around it
>may all still be in cache, so the second time gets a hot-cache boost). It
>may be useful to run the command once and *ignore* its times, and then
>re-run the command 3 times and average those results (so all 3 times you
>actually *use* start from the same "previous command just finished" cache state).
>  
>

Re: Why is Reiser4 slower then ReiserFS v3

[Adrian Ulrich]

I also noticed some odd slowness of reiser4

(Running 2.6.10 using the latest 2.6.10-rcsomething reiser4 patch)


What i did:

 I created a small script wich creates MANY (= 195075) directories
 like this: 1/[1-3]/[1-255]/[1-255]

 After this, i ran 'sync && find . > /dev/null && rm -rf *'

 Well, it's not a good test, but compare the speed of 
 the 'find' and 'rm -rf' part beteween ReiserFS and Reiser4 :-/

reiser3

#./mkdirs.pl
real    5m21.194s
user    1m12.323s
sys     4m6.740s

#find . > /dev/null
real    0m6.419s
user    0m0.940s
sys     0m5.159s

#rm -rf *
real    0m24.357s
user    0m1.558s
sys     0m22.797s

reiser4

#./mkdirs.pl
real    6m21.996s
user    1m13.395s
sys     5m8.233s

#find . > /dev/null
real    0m16.006s
user    0m3.323s
sys     0m12.676s

#rm -rf *
real    5m58.963s  <-- OUCH!
user    0m7.962s
sys     5m50.870s

Re: Why is Reiser4 slower then ReiserFS v3

[Hans Reiser]

Adrian Ulrich wrote:

>I also noticed some odd slowness of reiser4
>
>(Running 2.6.10 using the latest 2.6.10-rcsomething reiser4 patch)
>  
>
There was a read slowdown, in latest release of reiser4, see patch I 
cc'd this list on a few emails ago.

That said, try extents only option of reiser4 for faster deletes.

Because reiser4 has larger files stored in tails, it packs things more 
tightly than reiser3, but pays a price in delete speed to do it.  With 
extents only, no tight packing, fast deletes.

reiser3 also puts directory entries for multiple directories close to 
each other, and farther from the filebodies for them, compared to 
reiser4.  This may explain some of the find speed advantage.

Still, I am curious to see more measurements in this area if you have 
time for it.

Did you time the sync command or?  How large was the fileset created 
compared to RAM?

>
>What i did:
>
> I created a small script wich creates MANY (= 195075) directories
> like this: 1/[1-3]/[1-255]/[1-255]
>
> After this, i ran 'sync && find . > /dev/null && rm -rf *'
>
> Well, it's not a good test, but compare the speed of 
> the 'find' and 'rm -rf' part beteween ReiserFS and Reiser4 :-/
>
>reiser3
>
>#./mkdirs.pl
>real    5m21.194s
>user    1m12.323s
>sys     4m6.740s
>
>#find . > /dev/null
>real    0m6.419s
>user    0m0.940s
>sys     0m5.159s
>
>#rm -rf *
>real    0m24.357s
>user    0m1.558s
>sys     0m22.797s
>
>reiser4
>
>#./mkdirs.pl
>real    6m21.996s
>user    1m13.395s
>sys     5m8.233s
>
>#find . > /dev/null
>real    0m16.006s
>user    0m3.323s
>sys     0m12.676s
>
>#rm -rf *
>real    5m58.963s  <-- OUCH!
>user    0m7.962s
>sys     5m50.870s
>
>
>
>  
>

Re: Why is Reiser4 slower then ReiserFS v3

[Adrian Ulrich]

Hello Hans,

> There was a read slowdown, in latest release of reiser4, see patch I 
> cc'd this list on a few emails ago.

I saw the patch 5 seconds after i've posted my message ;)

I'll re-run my test using the patch and with/without the extents
option..

> Did you time the sync command or?

I ran a 'sync' after each test, but i didn't time it..
But i didn't feel a difference between the sync on
Reiser3 and Reiser4.. synching took less than 1/2 second on both
filesystems.


> How large was the fileset created compared to RAM?

I only created this 195075 directories, i've got about
1gb free RAM

When i re-run the test, i'll reboot after each command to get
an empty cache

Thanks,
 Adrian