View Single Post
Old 02-14-07, 03:50 PM   #3
floogy
Registered User
 
Join Date: Aug 2006
Posts: 34
Default Re: 2.6.20-rt5 and libata, sata_nv and ncq patch

Here is a thread I opened last year, but it's in german.
http://forum.ubuntuusers.de/topic/55202/

Quote:
Ein Benchmark unter xp x64 mit hd tach http://www.simplisoftware.com/Public...request=HdTach bringt es auf 95 Mb/s Durchsatz und 230Mb/s burst speed.
A benchmark under xp x64 with hd tach shows a throughput of 95 MB/sec and 230 MB burst speed.
I don't have raid enabled. I got only one hd.

With 2.6.20 the modul got parm adma set to true (I think that's maybe the whole patch, but also for ncq?):
Code:
modinfo sata_nv
filename:       /lib/modules/2.6.20-rt5/kernel/drivers/ata/sata_nv.ko
version:        3.2
license:        GPL
description:    low-level driver for NVIDIA nForce SATA controller
author:         NVIDIA
srcversion:     551FB8A96CE751C52091200
alias:          pci:v000010DEd*sv*sd*bc01sc04i*
alias:          pci:v000010DEd*sv*sd*bc01sc01i*
alias:          pci:v000010DEd000003F7sv*sd*bc*sc*i*
alias:          pci:v000010DEd000003F6sv*sd*bc*sc*i*
alias:          pci:v000010DEd000003E7sv*sd*bc*sc*i*
alias:          pci:v000010DEd0000037Fsv*sd*bc*sc*i*
alias:          pci:v000010DEd0000037Esv*sd*bc*sc*i*
alias:          pci:v000010DEd00000267sv*sd*bc*sc*i*
alias:          pci:v000010DEd00000266sv*sd*bc*sc*i*
alias:          pci:v000010DEd0000003Esv*sd*bc*sc*i*
alias:          pci:v000010DEd00000036sv*sd*bc*sc*i*
alias:          pci:v000010DEd00000055sv*sd*bc*sc*i*
alias:          pci:v000010DEd00000054sv*sd*bc*sc*i*
alias:          pci:v000010DEd000000EEsv*sd*bc*sc*i*
alias:          pci:v000010DEd000000E3sv*sd*bc*sc*i*
alias:          pci:v000010DEd0000008Esv*sd*bc*sc*i*
depends:        libata
vermagic:       2.6.20-rt5 SMP preempt mod_unload
parm:           adma:Enable use of ADMA (Default: true) (bool)

Benchmark:
Code:
  
dbench -t60 -D /tmp 50procs
dbench version 3.04 - Copyright Andrew Tridgell 1999-2004

Running for 60 seconds with load '/usr/share/dbench/client.txt' and minimum warmup 12 secs
50 clients started
  50        45    85.17 MB/sec  warmup   1 sec
  50        64    63.14 MB/sec  warmup   2 sec
  50        78    50.74 MB/sec  warmup   3 sec
  50       109    50.39 MB/sec  warmup   4 sec
  50       122    44.63 MB/sec  warmup   5 sec
  50       142    43.07 MB/sec  warmup   6 sec
  50       163    40.97 MB/sec  warmup   7 sec
  50       174    38.16 MB/sec  warmup   8 sec
  50       185    36.01 MB/sec  warmup   9 sec
  50       188    32.85 MB/sec  warmup  10 sec
  50       190    30.23 MB/sec  warmup  11 sec
  50       190    27.73 MB/sec  warmup  12 sec
  50       190    25.61 MB/sec  warmup  13 sec
  50       190    23.79 MB/sec  warmup  14 sec
  50       190    22.22 MB/sec  warmup  15 sec
  50       198    21.54 MB/sec  warmup  16 sec
  50       238    23.23 MB/sec  warmup  17 sec
  50       276    23.10 MB/sec  warmup  18 sec
  50       286    22.35 MB/sec  warmup  19 sec
  50       287    21.24 MB/sec  warmup  20 sec
  50       310    20.74 MB/sec  warmup  21 sec
  50       322    20.19 MB/sec  warmup  22 sec
  50       323    19.32 MB/sec  warmup  23 sec
  50       375    19.78 MB/sec  warmup  24 sec
  50       399    19.40 MB/sec  warmup  25 sec
  50       556    20.04 MB/sec  warmup  26 sec
  50       839    22.27 MB/sec  warmup  27 sec
  50      1040    23.55 MB/sec  warmup  28 sec
  50      1252    24.69 MB/sec  warmup  29 sec
  50      1491    35.49 MB/sec  execute   1 sec
  50      1637    42.56 MB/sec  execute   2 sec
  50      1672    35.06 MB/sec  execute   3 sec
[...]
  50      5851    44.45 MB/sec  execute  32 sec
  50      6611    48.65 MB/sec  execute  33 sec
  50      7340    53.28 MB/sec  execute  34 sec
  50      7634    57.25 MB/sec  execute  35 sec
  50      7637    55.67 MB/sec  execute  36 sec
  50      7647    54.26 MB/sec  execute  37 sec
  50      7650    52.84 MB/sec  execute  38 sec
  50      7673    51.63 MB/sec  execute  39 sec
  50      7724    50.72 MB/sec  execute  40 sec
[...]
  50      8236    28.26 MB/sec  cleanup  77 sec
  50      8236    27.89 MB/sec  cleanup  78 sec
  50      8236    27.75 MB/sec  cleanup  78 sec

Throughput 36.2631 MB/sec 50 procs
A second try:
Code:
 dbench -t60 -D /tmp 50procs
dbench version 3.04 - Copyright Andrew Tridgell 1999-2004

Running for 60 seconds with load '/usr/share/dbench/client.txt' and minimum warmup 12 secs
50 clients started
  50       885   206.61 MB/sec  warmup   1 sec
  50      1793   231.39 MB/sec  warmup   2 sec
  50      2049   233.44 MB/sec  warmup   3 sec
  50      2727   229.48 MB/sec  warmup   4 sec
  50      2800   193.57 MB/sec  warmup   5 sec
  50      2813   167.69 MB/sec  warmup   6 sec
  50      2824   147.06 MB/sec  warmup   7 sec
  50      2902   134.93 MB/sec  warmup   8 sec
  50      3690   144.11 MB/sec  warmup  10 sec
  50      3766   134.24 MB/sec  warmup  11 sec
  50      3828   125.13 MB/sec  warmup  12 sec
  50      4205    29.79 MB/sec  execute   1 sec
  50      4227    27.58 MB/sec  execute   2 sec
  50      4333    32.65 MB/sec  execute   3 sec
  50      4414    37.84 MB/sec  execute   4 sec
  50      4446    34.13 MB/sec  execute   5 sec
[...]
  50      9768    26.49 MB/sec  cleanup  71 sec
  50      9768    26.12 MB/sec  cleanup  72 sec
  50      9768    25.77 MB/sec  cleanup  73 sec
  50      9768    25.42 MB/sec  cleanup  74 sec
  50      9768    25.08 MB/sec  cleanup  75 sec
  50      9768    24.75 MB/sec  cleanup  76 sec
  50      9768    24.64 MB/sec  cleanup  76 sec

Throughput 31.3449 MB/sec 50 procs

Last year it came out that the service yacy wasn't configured right, what lead into heavy disk usage, and therefor poor benchmark performance.
floogy is offline   Reply With Quote