ATA disk performance (ICH2 controller), some tests andcomparison with Linux 2.6.5

[Mauro Triulzi]

Jeremy Chadwick wrote:

>A few things from my perspective:
>
>1)  Using dd to test disk performance is somewhat of a misnomer.  You
>should consider a tool like bonnie++ instead (ports/benchmarks/bonnie++).
>
>2)  Using block sizes of 512 bytes is pretty absurd.  Try other block
>sizes, preferably 8KB, 16KB, 32KB, and 64KB.  You may find disk
>performance is quite a bit better in such cases.
>
>3)  I happen to agree with Arne -- Linux is probably tricking you into
>believing the disk performance is higher than what it is due to memory
>caching.
>
>4)  There was a recent thread about ICH2-related issues (I believe
>performance was one of them) here on freebsd-current:
>
>http://lists.freebsd.org/mailman/htdig/freebsd-current/2004-September/036841.html
>
>5)  I have no idea what's up with the ATA channel re-init.  Soren should
>(and probably will) chime in here.
>
>6)  I'm still amazed people have systems in production that use ICH2.
>AFAIK, the chipset is deprecated.
>
>
Many thanks for your piece of advice. You and Arne got it right,
I have 755 MB Ram. I apologize for my naive approach.
Nevertheless I repeated the benchmarks this time with bonnie before and 
after
the ATA channel re-init:

i) Before re-init

foo# bonnie -s 1500
File './Bonnie.2561', size: 1572864000
Writing with putc()...done
Rewriting...done
Writing intelligently...done
Reading with getc()...done
Reading intelligently...done
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
              -------Sequential Output-------- ---Sequential Input-- 
--Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- 
--Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  
/sec %CPU
         1500 14533 12.8 14184  4.6  9360  3.2 27774 24.0 29067  5.7 
119.1  0.6

ii) ATA channell re-init

foo# atacontrol reinit 0
Master:  ad0 <IC35L060AVV207-0/V22OA63A> ATA/ATAPI revision 6
Slave:       no device present

and then

iii) After re-init

foo# bonnie -s 1500
File './Bonnie.2582', size: 1572864000
Writing with putc()...done
Rewriting...done
Writing intelligently...done
Reading with getc()...done
Reading intelligently...done
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
              -------Sequential Output-------- ---Sequential Input-- 
--Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- 
--Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  
/sec %CPU
         1500 27787 24.6 24690  8.3 12519  4.3 24487 21.4 24870  5.0 
128.1  0.6

iv) Bonnie in Linux 2.6.5

triulzi at linux:~/temp/bonnie> ./Bonnie  -s 1500
Bonnie 1.4: File './Bonnie.2732', size: 1572864000, volumes: 1
Writing with putc()...         done:  26859 kB/s  97.7 %CPU
Rewriting...                   done:  22480 kB/s   8.1 %CPU
Writing intelligently...       done:  59118 kB/s  14.0 %CPU
Reading with getc()...         done:  22508 kB/s  78.9 %CPU
Reading intelligently...       done:  48501 kB/s   9.0 %CPU
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
              ---Sequential Output (nosync)--- ---Sequential Input-- --Rnd 
Seek-
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k 
(03)-
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU   /sec 
%CPU
linux  1*1500 26859 97.7 59118 14.0 22480  8.1 22508 78.9 48501  9.0  177.1  
0.5

v) Comments

After re-init are the write and read performance per character in Linux
and FreeBSD very similar. Block-write, Block-read and rewrite are much
better in Linux. As Frode Nordahl rightly said are Linux and FreeBSD on 
different
parts of the disk and so it's not easy to compare I/O performance, but I 
don't think
that this explains such a difference.