Did anybody else notice this? The result of "hdparm -t" under 2.6.20-rc
seems to be less than half of what you get on 2.6.19. However, disk I/O
did *not* get slower according to bonnie++.
That is, there is no harm beyond confusion of the user. hdparm is lying
for some reason.
Below is the output of hdparm v6.6 and bonnie++ 1.93c. I ran hdparm 3
times each and bonnie++ 2 times to ensure that I get typical readings,
on an otherwise idle system.
hdparm -tT /dev/???:
========================================================================
2.6.19.1, IDE:
Timing cached reads: 10728 MB in 2.00 seconds = 5371.95 MB/sec
Timing buffered disk reads: 154 MB in 3.01 seconds = 51.11 MB/sec
2.6.19.1, SATA:
Timing cached reads: 8028 MB in 2.00 seconds = 4017.69 MB/sec
Timing buffered disk reads: 220 MB in 3.01 seconds = 73.18 MB/sec
2.6.19.1 + latest 1394 drivers, FireWire 800:
Timing cached reads: 10216 MB in 2.00 seconds = 5114.73 MB/sec
Timing buffered disk reads: 214 MB in 3.03 seconds = 70.71 MB/sec
2.6.19.1 + latest 1394 drivers, FireWire 400:
Timing cached reads: 8892 MB in 2.00 seconds = 4449.76 MB/sec
Timing buffered disk reads: 74 MB in 3.08 seconds = 24.02 MB/sec
========================================================================
2.6.20-rc3, IDE:
Timing cached reads: 11492 MB in 2.00 seconds = 5753.76 MB/sec
Timing buffered disk reads: 56 MB in 3.10 seconds = 18.09 MB/sec
2.6.20-rc3, SATA:
Timing cached reads: 10736 MB in 2.00 seconds = 5374.70 MB/sec
Timing buffered disk reads: 102 MB in 3.00 seconds = 33.99 MB/sec
2.6.20-rc3 + latest 1394 drivers, FireWire 800:
Timing cached reads: 9476 MB in 2.00 seconds = 4742.33 MB/sec
Timing buffered disk reads: 70 MB in 3.08 seconds = 22.69 MB/sec
2.6.20-rc3 + latest 1394 drivers, FireWire 400:
Timing cached reads: 10336 MB in 2.00 seconds = 5173.65 MB/sec
Timing buffered disk reads: 40 MB in 3.09 seconds = 12.95 MB/sec
========================================================================
bonnie++ on the SATA disk with a single 47% filled reiserfs partition:
========================================================================
2.6.19.1, SATA:
> Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
> Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
> stein 2G 203 99 57389 18 26943 9 856 93 66032 17 190.8 6
> Latency 48553us 295ms 1130ms 196ms 18075us 500ms
> Version 1.93c ------Sequential Create------ --------Random Create--------
> stein -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> 16 16541 55 +++++ +++ 18794 71 22933 79 +++++ +++ 24005 95
> Latency 107ms 129us 175ms 16613us 76us 236us
That is,
56.0 MB/s write and 64.5 MB/s read performance of sequential block I/O.
========================================================================
2.6.20-rc3, SATA:
> Version 1.93c ------Sequential Output------ --Sequential Input- --Random-
> Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
> stein 2G 210 99 57340 18 26920 10 1166 89 66805 17 195.3 5
> Latency 47794us 284ms 869ms 249ms 36229us 533ms
> Version 1.93c ------Sequential Create------ --------Random Create--------
> stein -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> 16 19645 67 +++++ +++ 21767 83 26951 94 +++++ +++ 24859 98
> Latency 231ms 85us 120us 469us 69us 118us
That is,
56.0 MB/s write and 65.2 MB/s read performance of sequential block I/O.
========================================================================
It's a 32bit kernel on a dual core x86 PC:
> $ uname -a
> Linux stein 2.6.20-rc3 #7 SMP PREEMPT Sat Jan 6 17:07:30 CET 2007 i686 Intel(R) Core(TM)2 CPU T7200 @ 2.00GHz GenuineIntel GNU/Linux
The motherboard is i945GT based.
> $ /usr/sbin/lspci
> 00:00.0 Host bridge: Intel Corporation Mobile Memory Controller Hub (rev 03)
> 00:01.0 PCI bridge: Intel Corporation Mobile PCI Express Graphics Port (rev 03)
> 00:02.0 VGA compatible controller: Intel Corporation Mobile Integrated Graphics Controller (rev 03)
> 00:1b.0 Class 0403: Intel Corporation 82801G (ICH7 Family) High Definition Audio Controller (rev 01)
> 00:1c.0 PCI bridge: Intel Corporation 82801G (ICH7 Family) PCI Express Port 1 (rev 01)
> 00:1c.4 PCI bridge: Intel Corporation 82801GR/GH/GHM (ICH7 Family) PCI Express Port 5 (rev 01)
> 00:1d.0 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #1 (rev 01)
> 00:1d.1 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #2 (rev 01)
> 00:1d.2 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #3 (rev 01)
> 00:1d.3 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #4 (rev 01)
> 00:1d.7 USB Controller: Intel Corporation 82801G (ICH7 Family) USB2 EHCI Controller (rev 01)
> 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev e1)
> 00:1f.0 ISA bridge: Intel Corporation 82801GB/GR (ICH7 Family) LPC Interface Bridge (rev 01)
> 00:1f.1 IDE interface: Intel Corporation 82801G (ICH7 Family) IDE Controller (rev 01)
> 00:1f.2 IDE interface: Intel Corporation 82801GB/GR/GH (ICH7 Family) Serial ATA Storage Controllers cc=IDE (rev 01)
> 00:1f.3 SMBus: Intel Corporation 82801G (ICH7 Family) SMBus Controller (rev 01)
> 01:00.0 PCI bridge: Texas Instruments Unknown device 8231 (rev 03)
> 02:02.0 FireWire (IEEE 1394): Texas Instruments TSB82AA2 IEEE-1394b Link Layer Controller (rev 01)
> 04:00.0 Ethernet controller: Intel Corporation 82573L Gigabit Ethernet Controller
> 05:04.0 FireWire (IEEE 1394): VIA Technologies, Inc. IEEE 1394 Host Controller (rev 80)
The IDE disk is a Hitachi HTE721010G9AT00 7200rpm 2.5".
The SATA disk is a Seagate ST3750640AS 7200rpm 3.5".
The FireWire disk is a dual ST3400832A 7200rpm 3.5" stripe-set on
OXFW921, tested on the 1394b and the 1394a controller.
Config is basically the same under both kernels:
http://me.in-berlin.de/~s5r6/linux1394/work-in-progress/other_stuff/config-diff-2.6.19.1--2.6.20-rc3-20070106.txt
http://me.in-berlin.de/~s5r6/linux1394/work-in-progress/other_stuff/config-2.6.20-rc3-20070106.txt
Userland is Gentoo 2006.1, kernel is Linus's with FireWire driver patches.
--
Stefan Richter
-=====-=-=== ---= --==-
http://arcgraph.de/sr/
On Sat, 6 Jan 2007, Stefan Richter wrote:
> Did anybody else notice this? The result of "hdparm -t" under 2.6.20-rc
> seems to be less than half of what you get on 2.6.19. However, disk I/O
> did *not* get slower according to bonnie++.
yes. See
http://lkml.org/lkml/2007/1/2/75
for the solution.
Tim
Tim Schmielau wrote:
> On Sat, 6 Jan 2007, Stefan Richter wrote:
>
>> Did anybody else notice this? The result of "hdparm -t" under 2.6.20-rc
>> seems to be less than half of what you get on 2.6.19. However, disk I/O
>> did *not* get slower according to bonnie++.
>
> yes. See
> http://lkml.org/lkml/2007/1/2/75
> for the solution.
>
> Tim
Thanks. I should have remembered that.
# hdparm -tT /dev/hda
/dev/hda:
Timing cached reads: 10864 MB in 2.00 seconds = 5440.10 MB/sec
Timing buffered disk reads: 58 MB in 3.06 seconds = 18.94 MB/sec
# cat /sys/block/hda/queue/scheduler
noop anticipatory deadline [cfq]
# echo anticipatory > /sys/block/hda/queue/scheduler
# hdparm -tT /dev/hda
/dev/hda:
Timing cached reads: 10680 MB in 2.00 seconds = 5347.20 MB/sec
Timing buffered disk reads: 148 MB in 3.02 seconds = 49.03 MB/sec
--
Stefan Richter
-=====-=-=== ---= --==-
http://arcgraph.de/sr/
> Tim Schmielau wrote:
>> See
>> http://lkml.org/lkml/2007/1/2/75
>> for the solution.
OK; this has already been committed a few days ago. I assume this closes
the issue. I will post after -rc4 in the unlikely case that it doesn't.
--
Stefan Richter
-=====-=-=== ---= --==-
http://arcgraph.de/sr/