slow sata hard drives in laptop

[Jakub Orlowski]

it seems that sata drives are being slow on laptops. i have had a few to play with and discovered slow read throughputs, i.e. 30MB/s maximum for at least 2 disks tested, which were hitachi made. there are a whole lot of websites reporting troubles with sata drives.
this should probably be worked over as it seems to affect boot time very much. (a dual core AMD turion x2 64bit booting as fast or even a bit slower than a intel celeron m 1,6GHz with 32 bit and only one core does not feel right)
it however seems not only to depend on the chipset, as this laptop here uses ICH7 and the other one was nvidia chipset.
maybe it could be helpful to see though the posts on the forums. there seems to be some trouble with the chipset driver, as some of the issues dissappear when enabling raid as the disk controller on bios. which will not happen on laptops due to limited bioses.

regards
jakub

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: linux-image-2.6.37-12-generic 2.6.37-12.26
Regression: Yes
ProcVersionSignature: Ubuntu 2.6.37-12.26-generic 2.6.37
Uname: Linux 2.6.37-12-generic i686
Architecture: i386
Date: Wed Jan 19 21:11:20 2011
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Alpha i386 (20100630.2)
ProcEnviron:
 LANGUAGE=de_DE:de:en_GB:en
 LANG=de_DE.UTF-8
 LC_MESSAGES=de_DE.utf8
 SHELL=/bin/bash
SourcePackage: linux
---
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.23.
Architecture: i386
ArecordDevices:
 **** List of CAPTURE Hardware Devices ****
 card 0: Intel [HDA Intel], device 0: AD198x Analog [AD198x Analog]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: jakub 1346 F.... pulseaudio
 /dev/snd/pcmC0D0p: jakub 1346 F...m pulseaudio
CRDA: Error: [Errno 2] Datei oder Verzeichnis nicht gefunden
Card0.Amixer.info:
 Card hw:0 'Intel'/'HDA Intel at 0xd0340000 irq 40'
   Mixer name : 'Analog Devices AD1986A'
   Components : 'HDA:11d41986,17aa2066,00100500'
   Controls : 20
   Simple ctrls : 11
DistroRelease: Ubuntu 11.04
Frequency: Once a day.
HibernationDevice: RESUME=UUID=8072d903-35ce-474e-8cdb-60bfee68c4b8
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Alpha i386 (20100630.2)
MachineType: LENOVO 076841G
Package: linux (not installed)
PccardctlIdent:
 Socket 0:
   no product info available
PccardctlStatus:
 Socket 0:
   no card
ProcEnviron:
 LANGUAGE=de_DE:de:en_GB:en
 LANG=de_DE.UTF-8
 LC_MESSAGES=de_DE.utf8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.37-12-generic root=UUID=6052c789-878e-40d9-8895-6485d758b4a1 ro vt.handoff=7 quiet splash
ProcVersionSignature: Ubuntu 2.6.37-12.26-generic 2.6.37
Regression: No
RelatedPackageVersions:
 linux-restricted-modules-2.6.37-12-generic N/A
 linux-backports-modules-2.6.37-12-generic N/A
 linux-firmware 1.45
Reproducible: No
Tags: natty running-unity kernel-uncat needs-upstream-testing
Uname: Linux 2.6.37-12-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
dmi.bios.date: 06/04/07
dmi.bios.vendor: LENOVO
dmi.bios.version: 61ET37WW
dmi.board.name: CAPELL VALLEY(NAPA) CRB
dmi.board.vendor: LENOVO
dmi.board.version: Not Applicable
dmi.chassis.type: 10
dmi.chassis.vendor: No Enclosure
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnLENOVO:bvr61ET37WW:bd06/04/07:svnLENOVO:pn076841G:pvr3000N100:rvnLENOVO:rnCAPELLVALLEY(NAPA)CRB:rvrNotApplicable:cvnNoEnclosure:ct10:cvrN/A:
dmi.product.name: 076841G
dmi.product.version: 3000 N100
dmi.sys.vendor: LENOVO

[Phillip Susi (psusi)]

Is this only with Natty, and works fine in Maverick? How did you test the throughput? Was it with hdparm -t? What does the manufacturer specify the throughput should be? Also please run apport-collect -p linux 705106 to collect more information.

[Jakub Orlowski (jakub-o)]

jakub@jakub-laptop:~$ dmesg|grep UDMA
[ 0.584958] ata1: SATA max UDMA/133 cmd 0x1f0 ctl 0x3f6 bmdma 0x18b0 irq 14
[ 0.584963] ata2: PATA max UDMA/100 cmd 0x170 ctl 0x376 bmdma 0x18b8 irq 15
[ 0.823346] ata2.00: ATAPI: HL-DT-ST DVDRAM GMA-4082N, HA01, max UDMA/33
[ 0.823542] ata1.00: ATA-7: HTS541080G9SA00, MB4IC60R, max UDMA/100
[ 0.836884] ata2.00: configured for UDMA/33
[ 0.837507] ata1.00: configured for UDMA/100
[ 1212.368271] ata2.00: configured for UDMA/33
[ 1214.612540] ata1.00: configured for UDMA/100
thats my dmesg including the name

jakub@jakub-laptop:~$ sudo hdparm -t /dev/sda
/dev/sda:
 Timing cached reads: 1644 MB in 2.00 seconds = 822.37 MB/sec
 Timing buffered disk reads: 92 MB in 3.06 seconds = 30.09 MB/sec up to 34MB

thats the specifications from hitachi's website
-Fast data transfer rate up to 1.5Gbps
- Media data transfer rate (max): 493 Mb/s
- Average seek time: 12 ms for read

umm which means that... ?

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Jakub Orlowski (jakub-o)]

apport information

[Phillip Susi (psusi)]

Based on that I would expect you to get more like 50 MB/s rater than 30. Still, I see nothing out of the ordinary or that indicates there is a bug. Unless you can do something, like run a different distribution or kernel, that would show it is not a hardware problem, I think it just is going to have to be chalked up to slow hardware.

[Jakub Orlowski (jakub-o)]

so laptop sata drives are slower than ide 3,5' ones for desktops in general?
if that is the case ... ok.
expecting it to be a driver thing, maybe i should have brought it up on kernel site, but i thought this would be the appropriate site to try first.
i am having this laptop since summer2010 so just ready for lucid maveric and now natty.
so its kernel 2.6.32-37
but i tried it on intrepid (2.6.28 i think), too. and it seemed quite fast. but i didnt benchmark there. if i find some old live cds and time i will try.
maybe someone can confirm that for different kernels.

[Jeremy Foshee (jeremyfoshee)]

Hi Jakub,

If you could also please test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Jakub Orlowski (jakub-o)]

ok... tested, with no change at all

palimpsest remains at
13 28 23 MB/s read throughput at 17,2ms access time. (hitachi wrote it to be ca. 60MB max and 12ms access)

and hdparm -tT /dev/sda:
/dev/sda:
 Timing cached reads: 1782 MB in 2.00 seconds = 891.55 MB/sec
 Timing buffered disk reads: 104 MB in 3.02 seconds = 34.44 MB/sec

i must say im confused though.
the drive connectors look like they are ide (parallel pins) - however thats only for the laptop i still have, which is the older one.
palimpsest gives mixed information:
it says pata host adapter, the driver is ata_piix but it says "SATA IDE Controller"
dmesg clearly says its SATA and there is a sticker on the drive telling it to be SATA.
the bios changed to SATA, but has been IDE before (firmware upgrade)

[Jakub Orlowski (jakub-o)]

jakub@jakub-laptop:~$ uname -r
2.6.38-999-generic
from
linux-image-2.6.38-999-generic_2.6.38-999.201101251128_i386.deb
and there was a message displayed justbefore plymouth set in:
[ 0.004333] /proc/device-tree: can't find root

i couldnt install the headers though. headers package says it needs the headers to install them...

and while talking about performance... why is ubuntu compiled with gcc 4.2.3? i thought the newer versions are much faster. or is it for debugging the upstream kernels only?
[ 0.000000] Linux version 2.6.38-999-generic (root@zinc) (gcc version 4.2.3 (Ubuntu 4.2.3-2ubuntu7)) #201101251128 SMP Tue Jan 25 12:39:33 UTC 2011

[Jakub Orlowski (jakub-o)]

and just one question concerning a different setup.
should software raid level 0 and 1 have the same read speeds? or level 1 better? if yes, then it is not the case.
(they have been made by alternative ubuntu cd installation - just in case there was any logic in there making raid 1 with different stripe size or whatever)

[Phillip Susi (psusi)]

Raid 0 is slightly faster at sequential reads, raid 1 is slightly faster at random.

[Jakub Orlowski (jakub-o)]

small update:
the connectors to my laptops HDD are Sata, i have been confused with another driver of a friend's laptop. So everything remains the same, and the "problem" persists.

[Launchpad Janitor (janitor)]

[Expired for linux (Ubuntu) because there has been no activity for 60 days.]

[Jakub Orlowski (jakub-o)]

expired huh?
i have just talked to friends of mine using Ubuntu reporting the same slow disk performance which do not occurr on MS Windows.
please someone... confirm it!