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

Sparc has slow disk access



Hello,

I´ve got a Sparc E250 server and I´m runing Debian GNU/Linux on it.
I now started to make some performance comparisons against a default PC.
Here are the configurations I used:

E250:
  ~> cat /proc/cpuinfo
  cpu             : TI UltraSparc II  (BlackBird)
  fpu             : UltraSparc II integrated FPU
  promlib         : Version 3 Revision 16
  prom            : 3.16.1
  type            : sun4u
  ncpus probed    : 2
  ncpus active    : 2
  Cpu0Bogo        : 799.53
  Cpu1Bogo        : 799.53
  MMU Type        : Spitfire
  State:
  CPU0:           online
  CPU1:           online

  2 GB RAM
  4x36 GB SCSI

  ~> uname -a
  Linux bse 2.2.18pre21 #1 SMP Tue Dec 19 22:53:19 CET 2000 sparc64 unknown

PC:
  ~> cat /proc/cpuinfo
  processor       : 0
  vendor_id       : GenuineIntel
  cpu family      : 6
  model           : 8
  model name      : Pentium III (Coppermine)
  stepping        : 6
  cpu MHz         : 795.576
  cache size      : 256 KB
  fpu             : yes
  fpu_exception   : yes
  cpuid level     : 2
  wp              : yes
  flags           : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr sse
  bogomips        : 1589.24

  256 MB RAM
  20 GB IDE

With this setup I noticed that the PC is about two times faster in
inserting data into a postgresql database.  I´ve got a hint to test
disk speed and I did so with the following script:


#!/bin/bash
date
ls -l /home/tillea/hd-test.file
i=0
while [ $i -lt 100 ] ; do
   echo -n "$i "
   let i="$i+1"
   dd if=/home/tillea/hd-test.file of=/tmp/hd-test.file 2> /dev/null
done
ls -l /tmp/hd-test.file
date
### end of scrip

Here are the results:

PC:

$ uname -a
Linux wr-linux02 2.4.5 #1 Mit Mai 30 08:24:06 CEST 2001 i686 unknown

$ hd-test
Don Aug 23 12:21:56 CEST 2001
-rw-r--r--    1 tillea   admin    25600000 23. Aug 11:50 /home/tillea/hd-test.file
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 -rw-r--r--    1 tillea   admin    25600000 23. Aug 12:22 /tmp/hd-test.file
Don Aug 23 12:22:47 CEST 2001

----> 51s  for the PC

Sparc:

$ uname -a
SunOS bse 5.8 Generic_108528-08 sun4u sparc SUNW,Ultra-250

$ hd-test
Thu Aug 23 13:23:38 MEST 2001
-rw-r--r--   1 tillea   admin    25600000 Aug 23 13:13 /home/tillea/hd-test.file
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 -rw-r--r--   1 tillea   admin    25600000 Aug 23 13:26 /tmp/hd-test.file
Thu Aug 23 13:26:49 MEST 2001

-----> 191s  for the Sparc running Solaris 8


$  uname -a
Linux bse 2.2.18pre21 #1 SMP Tue Dec 19 22:53:19 CET 2000 sparc64 unknown

Don Aug 23 13:55:37 CEST 2001
-rw-r--r--    1 tillea   admin    25600000 23. Aug 11:50 /home/tillea/hd-test.file
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 -rw-r--r--    1 tillea   admin    25600000 23. Aug 13:57 /tmp/hd-test.file
Don Aug 23 13:57:16 CEST 2001

-----> 99s    for tge Sparc running Debian GNU/Linux


So the relation is about:

   PC/linux    :      Sparc/Solaris     :      Sparc/Linux

      1        :          3.7           :         1.9

That´s not good for the Sparc machine (and even worse for Solaris by the way
but this is of less interest because the machine should run under Linux).

Any hint if there some optimisations could be done?

Kind regards

         Andreas.



Reply to: