All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Hancock <hancockr@shaw.ca>
To: "\"J.A. Magallón\"" <jamagallon@ono.com>
Cc: Linux-Kernel <linux-kernel@vger.kernel.org>
Subject: Re: iozone lies/bug ?
Date: Fri, 05 Sep 2008 19:23:17 -0600	[thread overview]
Message-ID: <48C1DB85.6020605@shaw.ca> (raw)
In-Reply-To: <fa.g18ZP/j70dwkyZRjfglzPcKZzdc@ifi.uio.no>

J.A. Magallón wrote:
> Hi all...
> 
> This is a curious thing, I think...
> I wanted to test the speed of an usb stick, and as a first step used hdparm
> on my root disk (real SCSI) and on the stick:
> 
> werewolf:/media/media# lsscsi
> [1:0:0:0]    disk    SEAGATE  ST336807LW       0C01  /dev/sda
> [1:0:1:0]    disk    SEAGATE  ST336807LW       0C01  /dev/sdb
> [2:0:0:0]    disk    ATA      ST3120022A       3.06  /dev/sdc
> [2:0:1:0]    cd/dvd  HL-DT-ST DVDRAM GSA-H10N  JL12  /dev/sr0
> [4:0:0:0]    disk    ATA      ST3200822AS      3.01  /dev/sdd
> [5:0:0:0]    disk    ATA      WDC WD3200AVJS-6 12.0  /dev/sde
> [7:0:0:0]    disk    USB      DISK 2.0         1219  /dev/sdf
> werewolf:/media/media# hdparm -tT /dev/sda /dev/sdf
> 
> /dev/sda:
>  Timing cached reads:   1402 MB in  2.00 seconds = 700.54 MB/sec
>  Timing buffered disk reads:  226 MB in  3.02 seconds =  74.86 MB/sec
> 
> /dev/sdf:
>  Timing cached reads:   1272 MB in  2.00 seconds = 635.69 MB/sec
>  Timing buffered disk reads:   52 MB in  3.02 seconds =  17.20 MB/sec
> 
> Well, looks reasonable.
> But, as hdparm is not a 'serious' test, and I wanted also write speed,
> used iozone:
> 
> werewolf:/media/media# iozone -v
>        'Iozone' Filesystem Benchmark Program
> 
>         Version $Revision: 3.287 $
>     Compiled for 32 bit mode.
> 
> But it gives strange results:
> 
> for sda1 (/):
> 
>     Machine = Linux werewolf.home 2.6.26-jam07 #1 SMP PREEMPT Mon Sep 1 01:47:23 C  Auto Mode
>     Auto Mode
>     File size set to 4096 KB
>     Command line used: iozone -M -a -s 4m
>     Output is in Kbytes/sec
>     Time Resolution = 0.000001 seconds.
>     Processor cache size set to 1024 Kbytes.
>     Processor cache line size set to 32 bytes.
>     File stride size set to 17 * record size.
>                                                             random  random    bkwd  record  stride
>               KB  reclen   write rewrite    read    reread    read   write    read rewrite    read   fwrite frewrite   fread  freread
>             4096       4  180855  742576  1275252  1130840 1143176  597607  998780 670266 1168616   231962   670815 1071129  1259145
>             4096       8  247673  809183  1246875  1407083 1351815  711728 1176673 854936 1298668   236817   744321 1233363  1519851
>             4096      16  251644  851010  1123131  1691931 1668431  716459 1094017 931544 1259145   245622   793337 1110328  1659643
>             4096      32  253245  871663  1463432  1764719 1808388  757396 1339875 885813 1393197   261107   796112 1128374  1742978
>             4096      64  254837  862332  1139037  1789405 1585752  779448 1136199 961728 1844214   259684   799687 1124965  1538114
>             4096     128  250090  868154  1177029  1715966 1700996  771810 1181084 980842 1643659   228405   792569 1003429  1720285
>             4096     256  242984  791959   950372   718080  930274  826472  694472 708038  913470   223751   566606  698737   918385
>             4096     512  238696  664406   764459   803304  720492  519335  727660 420102  686557   212735   416768  728695   695534
>             4096    1024  232926  667198   711975   773266  735896  689562  781828 663857  742297   253183   541297  658732   816912
>             4096    2048  243548  659167   653899   738807  728048  632684  638304 493137  599707   305011   660858  664180   792569
>             4096    4096  218417  655142   857449   732729  694826  696717  689098 557278  627643   242913   610705  713588   784073
> 
> for sdf1:
> 
>     Machine = Linux werewolf.home 2.6.26-jam07 #1 SMP PREEMPT Mon Sep 1 01:47:23 C  Auto Mode
>     Auto Mode
>     File size set to 4096 KB
>     Command line used: iozone -M -a -s 4m
>     Output is in Kbytes/sec
>     Time Resolution = 0.000001 seconds.
>     Processor cache size set to 1024 Kbytes.
>     Processor cache line size set to 32 bytes.
>     File stride size set to 17 * record size.
>                                                             random  random    bkwd  record  stride
>               KB  reclen   write rewrite    read    reread    read   write    read rewrite    read   fwrite frewrite   fread  freread
>             4096       4  391850  843302   885058  1159992  746900  704385  778707 566371  941825   481712   772247 1204705   795494
>             4096       8  385690  981087  1036709  1419270  945740  755162  978967 555842 1211475   390392   919622 1209329  1071970
>             4096      16  613438 1121897  1335497  1594434 1441745  683008 1217598 973152 1117903   558038  1014363 1144453  1552103
>             4096      32  559635  968335  1494912  1601847 1676627 1129619 1491081 1264978 1675941   604932  1051334 1217598  1629276
>             4096      64  627925 1197687  1557477  1697406 1573569 1082738 1589445 1237462 1755679   587830  1090812 1216513  1678000
>             4096     128  538446 1177653  1366266  1268879  963537  796422  981548 983197 1582077   428272   977099 1431167  1643000
>             4096     256  611078 1160652   907177  1041725 1114861 1214349 1088204 1161985  813343   454051   552021  919003   984852
>             4096     512  448633  912842   725981   782419  722526  983197  579103 870748  644937   432341   632684  718092   768336
>             4096    1024  473859  816750   717209   795504  738683  660964  694472 793337  716083   385070   699573  718092   760207
>             4096    2048  513161  789051   685059   768041  718974  652437  643418 639500  715583   486749   881049  715708   747854
>             4096    4096  505174  700760   835748   775173  786029  794414  797507 718470  594830   408904   835066  637608   670815
> 
> First, Output is not in Kb, or I have a disk that gives 1Gb/s... :).
> And second, how can both drives give the same ? The file is too small ?
> Do I need -e to iozone to account for sync ? But that would not affect read
> speed...
> Strange (for my understanding).

A file size of 4096KB is probably just running entirely out of the cache..

       reply	other threads:[~2008-09-06  1:23 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <fa.g18ZP/j70dwkyZRjfglzPcKZzdc@ifi.uio.no>
2008-09-06  1:23 ` Robert Hancock [this message]
2008-09-05 17:03 iozone lies/bug ? J.A. Magallón

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=48C1DB85.6020605@shaw.ca \
    --to=hancockr@shaw.ca \
    --cc=jamagallon@ono.com \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.