Re: IDE Harddrive Performance

Re: IDE Harddrive Performance

[Brendan Pike]

On Thursday 20 December 2001 02:08 pm, you wrote:
> > here is -T
> >
> > [root@spikes spike]# /sbin/hdparm -T /dev/hda
> > /dev/hda:
> > Timing buffer-cache reads:   128 MB in  3.89 seconds = 32.90 MB/sec
>
> well, no wonder.  on a modern machine, -T shows 200 MB/s.
> what kind of CPU/motherboard is this?  I'm guessing it's
> a socket7 system with PC66 dram, perhaps just edo.  the problem
> is that it's dram and probably PCI infrastructure can't
> really handle much bandwidth, so doesn't keep the disk "streaming".

Actually it is a Socket 7 AMD 300 with 384MB SDRAM PC-133 running at 100Mhz 
Bus clock.

Re: IDE Harddrive Performance

[Wayne Whitney]

Jim and Adam,

I recommend that you check out the smartsuite package described in
http://www.linux-ide.org/smart.html.  It showed me that when a Maxtor
drive I recently purchased dropped from 40MB/sec to 1MB/sec throughput, it
was having an incredible number of "Hardware ECC Recovered" (SMART
attribute 195) events.  I guess some bit chunk of the magnetic media had
failed, in a way that still allowed the drive to recover all the data, but
so that the overhead of the requisite ECC was killing the throughput.  I
ended up returning it for a replacement.

Cheers, Wayne

Re: IDE Harddrive Performance

[Mark Hahn]

> I recommend that you check out the smartsuite package described in
> http://www.linux-ide.org/smart.html.  It showed me that when a Maxtor
> drive I recently purchased dropped from 40MB/sec to 1MB/sec throughput, it
> was having an incredible number of "Hardware ECC Recovered" (SMART
> attribute 195) events.  I guess some bit chunk of the magnetic media had

the smart tools are *VERY* useful.  I recently built a raid box and
installed 5 of the disks in one nice fan-equipped bracket.  unfortunately,
100G WD 7200 RPM's vibrate fairly seriously, and this actually caused
horrible performance, as well as a small number of bad low-level writes 
and thus remapped sectors (shown by smartctl).  simply moving some of the
disks elsewhere entirely fixed the problem (120 MB/s!)

regards, mark hahn.

Re: IDE Harddrive Performance

[Jim Radford]

> hda: Maxtor 93073U6, ATA DISK drive
> hdc: Maxtor 90840D6, ATA DISK drive
 
> /dev/hda:
>  Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec
> /dev/hdc:
>  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec

I'm in the same boat with my Maxtor 54098U8.  It has been getting
progressively slower.  Two weeks ago I was getting a whopping 3MB/s,
now I get:

/dev/hda:
 Timing buffered disk reads:  64 MB in 115.29 seconds =568.44 kB/sec

My CDROM (ide-scsi) is faster than this!

# time dd if=/dev/scd0 of=/dev/null count=32k bs=512
real    0m8.443s

  = 1.8MB/sec

The wierd thing is that I don't get any errors, but I have just
switched from emacs to vi. :-) I ran Maxtor's powermax.exe and their
90sec test hadn't finished after 20 minutes so I rebooted. :-(

Same results with 2.4.9 - 2.4.17.

-Jim

VP_IDE: VIA vt82c686a (rev 14) IDE UDMA66 controller on pci00:07.1

# hdparm /dev/hda

/dev/hda:
 multcount    = 16 (on)
 I/O support  =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 4865/255/63, sectors = 78165360, start = 0
 busstate     =  1 (on)

Re: IDE Harddrive Performance

[Craig Knox]

On Sat, 2001-12-22 at 18:20, Jim Radford wrote:
> > hda: Maxtor 93073U6, ATA DISK drive
> > hdc: Maxtor 90840D6, ATA DISK drive
>  
> > /dev/hda:
> >  Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec
> > /dev/hdc:
> >  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
> 
> I'm in the same boat with my Maxtor 54098U8.  It has been getting
> /dev/hda:
>  Timing buffered disk reads:  64 MB in 115.29 seconds =568.44 kB/sec
> Same results with 2.4.9 - 2.4.17.
> VP_IDE: VIA vt82c686a (rev 14) IDE UDMA66 controller on pci00:07.1
> 
> # hdparm /dev/hda
> 
> /dev/hda:
>  multcount    = 16 (on)
>  I/O support  =  1 (32-bit)
>  unmaskirq    =  1 (on)
>  using_dma    =  1 (on)
>  keepsettings =  0 (off)
>  nowerr       =  0 (off)
>  readonly     =  0 (off)
>  readahead    =  8 (on)
>  geometry     = 4865/255/63, sectors = 78165360, start = 0
>  busstate     =  1 (on)

Haven't had any trouble 2.4.x - 2.4.17 using the same model drive.
PDC20268: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER 
hde: Maxtor 54098U8, ATA DISK drive
hdg: Maxtor 54098U8, ATA DISK drive

with 2.4.17 I get ~28.96MB/sec with hdparm -t /dev/hdx

# hdparm /dev/hde
multicount	= 16 (on)
I/O support	=  3 (32bit w/sync)
unmaskirq	=  0 (off)
using_dma	=  1 (on)
keepsettings	=  0 (off)
nowerr		=  0 (off)
readonly	=  0 (off)
readahead	=  8 (on)
geometry	= 4111/255/63, sectors = 66055248, start = 0

Re: IDE Harddrive Performance

[Adam Keys]

On December 22, 2001 12:20, Jim Radford wrote:
> > hda: Maxtor 93073U6, ATA DISK drive
> > hdc: Maxtor 90840D6, ATA DISK drive
> >
> > /dev/hda:
> >  Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec
> > /dev/hdc:
> >  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
>
> I'm in the same boat with my Maxtor 54098U8.  It has been getting
> progressively slower.  Two weeks ago I was getting a whopping 3MB/s,
> now I get:
>
> /dev/hda:
>  Timing buffered disk reads:  64 MB in 115.29 seconds =568.44 kB/sec

I just ripped my machine apart, cleaned it out and made sure all the 
connections are snug.  I didn't find anything glaringly wrong inside, so my 
current conclusion is that my machine just gets too durn hot for the drive to 
function properly.  I'm going to periodically test my throughput to see if it 
degrades over time.  See if taking your machine down for a few hours helps 
things out.  If so we can yell at Maxtor, otherwise its back on the bug hunt.

Thanks to everyone who sent suggestions!  Hopefully I'm through with this 
hiccup. :)
-- 
akk~

Re: IDE Harddrive Performance

[Brendan Pike]

Errk, yeah the system was under a high load at the time i did that, 90+% cpu.
I now remember doing it a few days ago and it was about 140 MB/s.
The system is under load right now so i cant get another reliable reading.

Re: IDE Harddrive Performance

[Brendan Pike]

here is -T

[root@spikes spike]# /sbin/hdparm -T /dev/hda 
/dev/hda:
Timing buffer-cache reads:   128 MB in  3.89 seconds = 32.90 MB/sec

Re: IDE Harddrive Performance

[Ryan Cumming]

On December 20, 2001 09:10, Brendan Pike wrote:
> [root@spikes spike]# /sbin/hdparm -T /dev/hda
> /dev/hda:
> Timing buffer-cache reads:   128 MB in  3.89 seconds = 32.90 MB/sec

phalynx:/home/bodnar42# hdparm -T /dev/hda

/dev/hda:
 Timing buffer-cache reads:   128 MB in  0.63 seconds =203.17 MB/sec

Either you have an old computer with absolutely horrible memory bandwidth, or 
something is really wrong here.

-Ryan

Re: IDE Harddrive Performance

[Sebastian Dröge]

[-- Attachment #1: Type: text/plain, Size: 5288 bytes --]

Hi,
My disc throughput on a Intel PIIX4 with an IBM-DTTA-351010 (10 GB/hda) and a Western Digital WD800BB-00BSA0 (80 GB/hdb)
Kernel is 2.5.1-dj2
Maybe it helps a bit ;)
Little question for the masters of hardware here in the list... buffer-chache reads speed is just the speed of the cache, right?
So I think only buffered disk reads say something about the real speed of the drive. And 10 or 11 MB/sec aren't much...

---------------------------------------------------------------------------------------------------

root:/home/slomo# hdparm -Tt /dev/hda

/dev/hda:
 Timing buffer-cache reads:   128 MB in  1.49 seconds = 85.91 MB/sec
 Timing buffered disk reads:  64 MB in  5.93 seconds = 10.79 MB/sec

root:/home/slomo# hdparm /dev/hda

/dev/hda:
 multcount    = 16 (on)
 I/O support  =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 1232/255/63, sectors = 19807200, start = 0

root:/home/slomo# hdparm -I /dev/hda

/dev/hda:

non-removable ATA device, with non-removable media
        Model Number:           IBM-DTTA-351010
        Serial Number:                   WF0WFHH5554
        Firmware Revision:      T56OA73A
Standards:
        Used: ATA/ATAPI-4 T13 1153D revision 17
        Supported: 1 2 3 4
Configuration:
        Logical         max     current
        cylinders       16383   16383
        heads           16      16
        sectors/track   63      63
        bytes/track:    0               (obsolete)
        bytes/sector:   0               (obsolete)
        current sector capacity: 16514064
        LBA user addressable sectors = 19807200
Capabilities:
        LBA, IORDY(can be disabled)
        Buffer size: 466.0kB    ECC bytes: 34   Queue depth: 32
        Standby timer values: spec'd by standard, no device specific minimum
        r/w multiple sector transfer: Max = 16  Current = 16
        DMA: sdma0 sdma1 sdma2 mdma0 mdma1 mdma2 udma0 udma1 *udma2
             Cycle time: min=120ns recommended=120ns
        PIO: pio0 pio1 pio2 pio3 pio4
             Cycle time: no flow control=240ns  IORDY flow control=120ns
Commands/features:
        Enabled Supported:
           *    NOP cmd
           *    READ BUFFER cmd
           *    WRITE BUFFER cmd
           *    Host Protected Area feature set
                release interrupt
           *    look-ahead
           *    write cache
           *    Power Management feature set
                Security Mode feature set
           *    SMART feature set
           *    READ/WRITE DMA QUEUED
Security:
                supported
        not     enabled
        not     locked
        not     frozen
        not     expired: security count
        not     supported: enhanced erase
        22min for SECURITY ERASE UNIT.

-------------------------------------------------------------------------------------------------

root:/home/slomo# hdparm -Tt /dev/hdb

/dev/hdb:
 Timing buffer-cache reads:   128 MB in  1.39 seconds = 92.09 MB/sec
 Timing buffered disk reads:  64 MB in  5.61 seconds = 11.41 MB/sec

root:/home/slomo# hdparm /dev/hdb

/dev/hdb:
 multcount    = 16 (on)
 I/O support  =  1 (32-bit)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 155061/16/63, sectors = 156301488, start = 0

root:/home/slomo# hdparm -I /dev/hdb

/dev/hdb:

non-removable ATA device, with non-removable media
        Model Number:           WDC WD800BB-00BSA0
        Serial Number:          WD-WMA6S1226085
        Firmware Revision:      12.08C12
Standards:
        Supported: 1 2 3 4 5
        Likely used: 5
Configuration:
        Logical         max     current
        cylinders       16383   17475
        heads           16      15
        sectors/track   63      63
        bytes/track:    57600           (obsolete)
        bytes/sector:   600             (obsolete)
        current sector capacity: 16513875
        LBA user addressable sectors = 156301488
Capabilities:
        LBA, IORDY(can be disabled)
        Buffer size: 2048.0kB   ECC bytes: 40   Queue depth: 1
        Standby timer values: spec'd by standard, with device specific minimum
        r/w multiple sector transfer: Max = 16  Current = 16
        DMA: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4 udma5
             Cycle time: min=120ns recommended=120ns
        PIO: pio0 pio1 pio2 pio3 pio4
             Cycle time: no flow control=120ns  IORDY flow control=120ns
Commands/features:
        Enabled Supported:
           *    READ BUFFER cmd
           *    WRITE BUFFER cmd
           *    Host Protected Area feature set
           *    look-ahead
           *    write cache
           *    Power Management feature set
                Security Mode feature set
           *    SMART feature set
                SET MAX security extension
           *    DOWNLOAD MICROCODE cmd
Security:
                supported
        not     enabled
        not     locked
        not     frozen
        not     expired: security count
        not     supported: enhanced erase
HW reset results:
        CBLID- above Vih
        Device num = 1 determined by the jumper
Checksum: correct


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: IDE Harddrive Performance

[Brendan Pike]

Well i really cant explain this,

[root@spikes spike]# /sbin/hdparm -t /dev/hda
/dev/hda:
Timing buffered disk reads:  64 MB in  6.65 seconds =  9.62 MB/sec

[root@spikes spike]# cat /proc/ide/via
----------VIA BusMastering IDE Configuration----------------
Driver Version:                     3.29
South Bridge:                       VIA vt82c586b
Revision:                           ISA 0x47 IDE 0x6
Highest DMA rate:                   UDMA33
BM-DMA base:                        0xe000
PCI clock:                          33MHz
Master Read  Cycle IRDY:            1ws
Master Write Cycle IRDY:            1ws
BM IDE Status Register Read Retry:  yes
Max DRDY Pulse Width:               No limit
-----------------------Primary IDE-------Secondary IDE------
Read DMA FIFO flush:          yes                 yes
End Sector FIFO flush:         no                  no
Prefetch Buffer:              yes                 yes
Post Write Buffer:            yes                  no
Enabled:                      yes                 yes
Simplex only:                  no                  no
Cable Type:                   40w                 40w
-------------------drive0----drive1----drive2----drive3-----
Transfer Mode:       UDMA       PIO      UDMA       PIO
Address Setup:       30ns     120ns      30ns     120ns
Cmd Active:          90ns      90ns      90ns      90ns
Cmd Recovery:        30ns      30ns      30ns      30ns
Data Active:         90ns     330ns      90ns     330ns
Data Recovery:       30ns     270ns      30ns     270ns
Cycle Time:          60ns     600ns      60ns     600ns
Transfer Rate:   33.0MB/s   3.3MB/s  33.0MB/s   3.3MB/s

[root@spikes spike]# /sbin/hdparm -i /dev/hda

/dev/hda:

 Model=FUJITSU MPG3409AT E, FwRev=82B9, SerialNo=VH06T1408HST
 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
 RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
 BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
 CurCHS=16383/16/63, CurSects=-66060037, LBA=yes, LBAsects=80063424
 IORDY=yes, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes: pio0 pio1 pio2 pio3 pio4
 DMA modes: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4 udma5

[root@spikes spike]# /sbin/hdparm /dev/hda

/dev/hda:
 multcount    = 16 (on)
 I/O support  =  3 (32-bit w/sync)
 unmaskirq    =  1 (on)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 4983/255/63, sectors = 80063424, start = 0


Kernel 2.4.16 without patches.

IDE Harddrive Performance

[Thomas Deselaers]

Hello,

I do have an Asus P2B-S Mainboard and since a week I have a Maxtor 60 GB
5400 rpm Harddrive (MAXTOR 4K060H3).

I tried the performance of the drive and got some results which are quite
low I think.

hdparm -t /dev/hdc returns

/dev/hdc:
 Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
 
What would be a value I can expect from my hardware? And what might result
in higher speeds?

thanks,
thomas
-- 
thomas@deselaers.de «<>» JabberID on request «<>» GPG/PGP key on request
  «< Unless stated otherwise everything I write is just my opinion >»

Re: IDE Harddrive Performance

[Brendan Pike]

On Wednesday 19 December 2001 11:32 am, Thomas Deselaers wrote:
> Hello,
>
> I do have an Asus P2B-S Mainboard and since a week I have a Maxtor 60 GB
> 5400 rpm Harddrive (MAXTOR 4K060H3).
>
> I tried the performance of the drive and got some results which are quite
> low I think.
>
> hdparm -t /dev/hdc returns
>
> /dev/hdc:
>  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
>
> What would be a value I can expect from my hardware? And what might result
> in higher speeds?
>
> thanks,
> thomas

I dont really know, I dont think its possible to get higher then that from a 
5400 RPM disk. Heres mine,

FUJITSU 40.9GB MPG3409AT 5400 RPM

/dev/hda:
Timing buffered disk reads:  64 MB in  7.96 seconds =  8.04 MB/sec
/dev/hda:
Timing buffered disk reads:  64 MB in  7.07 seconds =  9.05 MB/sec
/dev/hda:
Timing buffered disk reads:  64 MB in  6.70 seconds =  9.55 MB/sec

Seems pretty normal to me. However yours is alot better then mine.

Re: IDE Harddrive Performance

[Lee Packham]

Dunno, just tried this on my box which has a 5400 RPM disk in and got 

[root@smeg lpackham]# /sbin/hdparm -t /dev/hdc

7200RPM on the on board
/dev/hdc:
 Timing buffered disk reads:  64 MB in  3.94 seconds = 16.24 MB/sec
[root@smeg lpackham]# /sbin/hdparm -t /dev/hdg

7200RPM on a promise Ultra ATA 66
/dev/hdg:
 Timing buffered disk reads:  64 MB in  3.02 seconds = 21.19 MB/sec
[root@smeg lpackham]# /sbin/hdparm -t /dev/hda

5400RPM on the on board
/dev/hda:
 Timing buffered disk reads:  64 MB in  4.23 seconds = 15.13 MB/sec


> On Wednesday 19 December 2001 11:32 am, Thomas Deselaers wrote:
>> Hello,
>>
>> I do have an Asus P2B-S Mainboard and since a week I have a Maxtor 60
>> GB 5400 rpm Harddrive (MAXTOR 4K060H3).
>>
>> I tried the performance of the drive and got some results which are
>> quite low I think.
>>
>> hdparm -t /dev/hdc returns
>>
>> /dev/hdc:
>>  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
>>
>> What would be a value I can expect from my hardware? And what might
>> result in higher speeds?
>>
>> thanks,
>> thomas
> 
> I dont really know, I dont think its possible to get higher then that
> from a  5400 RPM disk. Heres mine,
> 
> FUJITSU 40.9GB MPG3409AT 5400 RPM
> 
> /dev/hda:
> Timing buffered disk reads:  64 MB in  7.96 seconds =  8.04 MB/sec
> /dev/hda:
> Timing buffered disk reads:  64 MB in  7.07 seconds =  9.05 MB/sec
> /dev/hda:
> Timing buffered disk reads:  64 MB in  6.70 seconds =  9.55 MB/sec
> 
> Seems pretty normal to me. However yours is alot better then mine. -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
> in the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: IDE Harddrive Performance

[Jan Niehusmann]

On Wed, Dec 19, 2001 at 11:44:47AM -0400, Brendan Pike wrote:
> I dont really know, I dont think its possible to get higher then that from a 
> 5400 RPM disk. Heres mine,

/dev/hda:
 Timing buffered disk reads:  64 MB in  2.32 seconds = 27.59 MB/sec

bash-2.05a# cat /proc/ide/hda/model 
Maxtor 98196H8

This is a 5400rpm drive, too.

Jan

Re: IDE Harddrive Performance

[Brendan Pike]

On Wednesday 19 December 2001 12:01 pm, Jan Niehusmann wrote:
> On Wed, Dec 19, 2001 at 11:44:47AM -0400, Brendan Pike wrote:
> > I dont really know, I dont think its possible to get higher then that
> > from a 5400 RPM disk. Heres mine,
>
> /dev/hda:
>  Timing buffered disk reads:  64 MB in  2.32 seconds = 27.59 MB/sec
>
> bash-2.05a# cat /proc/ide/hda/model
> Maxtor 98196H8
>
> This is a 5400rpm drive, too.
>
> Jan
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

wowie, that is quite slow then. well udma33 mode is probebly why. if getting 
an ata100 offboard card (since i have an ata100 drive) would make such a big 
differance, im all for it. would there be any others reasons for such 
slowness? is udma33 capible of more then 9MB/sec ??

[root@spikes spike]# cat /proc/ide/via
----------VIA BusMastering IDE Configuration----------------
Driver Version:                     3.29
South Bridge:                       VIA vt82c586b
Revision:                           ISA 0x47 IDE 0x6
Highest DMA rate:                   UDMA33
BM-DMA base:                        0xe000
PCI clock:                          33MHz
Master Read  Cycle IRDY:            1ws
Master Write Cycle IRDY:            1ws
BM IDE Status Register Read Retry:  yes
Max DRDY Pulse Width:               No limit
-----------------------Primary IDE-------Secondary IDE------
Read DMA FIFO flush:          yes                 yes
End Sector FIFO flush:         no                  no
Prefetch Buffer:              yes                 yes
Post Write Buffer:            yes                  no
Enabled:                      yes                 yes
Simplex only:                  no                  no
Cable Type:                   40w                 40w
-------------------drive0----drive1----drive2----drive3-----
Transfer Mode:       UDMA       PIO       PIO       PIO
Address Setup:       30ns     120ns      30ns     120ns
Cmd Active:          90ns      90ns      90ns      90ns
Cmd Recovery:        30ns      30ns      30ns      30ns
Data Active:         90ns     330ns      90ns     330ns
Data Recovery:       30ns     270ns      30ns     270ns
Cycle Time:          90ns     600ns     120ns     600ns
Transfer Rate:   22.0MB/s   3.3MB/s  16.5MB/s   3.3MB/s

dont know if that info would help much but there it is. also uses the via 
kernel driver.

Re: IDE Harddrive Performance

[Oleg Artamonov]

Brendan Pike написал:

> wowie, that is quite slow then. well udma33 mode is probebly why. if
> getting an ata100 offboard card (since i have an ata100 drive) would make
> such a big differance, im all for it. would there be any others reasons for
> such slowness? is udma33 capible of more then 9MB/sec ??

  Yes, of course it's capable. Up to 33Mbytes/sec (really you should see 
about 20Mbytes/sec or more). Sounds like you disable 32-bit transfer or 
disable DMA modes.

> dont know if that info would help much but there it is. also uses the via
> kernel driver.

  What 'hdparm /dev/<your drive>' and 'hdparm -i /dev/<your drive>' says?

Re: IDE Harddrive Performance

[Vojtech Pavlik]

On Wed, Dec 19, 2001 at 12:18:13PM -0400, Brendan Pike wrote:
> On Wednesday 19 December 2001 12:01 pm, Jan Niehusmann wrote:
> > On Wed, Dec 19, 2001 at 11:44:47AM -0400, Brendan Pike wrote:
> > > I dont really know, I dont think its possible to get higher then that
> > > from a 5400 RPM disk. Heres mine,
> >
> > /dev/hda:
> >  Timing buffered disk reads:  64 MB in  2.32 seconds = 27.59 MB/sec
> >
> > bash-2.05a# cat /proc/ide/hda/model
> > Maxtor 98196H8
> >
> > This is a 5400rpm drive, too.
> >
> > Jan
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> 
> wowie, that is quite slow then. well udma33 mode is probebly why. if getting 
> an ata100 offboard card (since i have an ata100 drive) would make such a big 
> differance, im all for it. would there be any others reasons for such 
> slowness? is udma33 capible of more then 9MB/sec ??

Yes. Reaching up to 16 MB/sec with udma33 is quite possible.

> [root@spikes spike]# cat /proc/ide/via
> ----------VIA BusMastering IDE Configuration----------------
> Driver Version:                     3.29
> South Bridge:                       VIA vt82c586b
> Revision:                           ISA 0x47 IDE 0x6
> Highest DMA rate:                   UDMA33
> BM-DMA base:                        0xe000
> PCI clock:                          33MHz
> Master Read  Cycle IRDY:            1ws
> Master Write Cycle IRDY:            1ws
> BM IDE Status Register Read Retry:  yes
> Max DRDY Pulse Width:               No limit
> -----------------------Primary IDE-------Secondary IDE------
> Read DMA FIFO flush:          yes                 yes
> End Sector FIFO flush:         no                  no
> Prefetch Buffer:              yes                 yes
> Post Write Buffer:            yes                  no
> Enabled:                      yes                 yes
> Simplex only:                  no                  no
> Cable Type:                   40w                 40w
> -------------------drive0----drive1----drive2----drive3-----
> Transfer Mode:       UDMA       PIO       PIO       PIO
> Address Setup:       30ns     120ns      30ns     120ns
> Cmd Active:          90ns      90ns      90ns      90ns
> Cmd Recovery:        30ns      30ns      30ns      30ns
> Data Active:         90ns     330ns      90ns     330ns
> Data Recovery:       30ns     270ns      30ns     270ns
> Cycle Time:          90ns     600ns     120ns     600ns
> Transfer Rate:   22.0MB/s   3.3MB/s  16.5MB/s   3.3MB/s
> 
> dont know if that info would help much but there it is. also uses the via 
> kernel driver.

But for some reason your drive is running at udma22 (udma1). 

-- 
Vojtech Pavlik
SuSE Labs

Re: IDE Harddrive Performance

[Dmitry Pogosyan]

On Asus P2B and  QUANTUM FIREBALLP AS10.2  10GB 5400 RPM drive in udma2
mode
I'm getting  ~17 MB/s

            Dmitri

Re: IDE Harddrive Performance

[Thomas Deselaers]

On Wed, Dec 19, 2001 at 11:20:22AM -0500, Mark Hahn wrote:
> > > I do have an Asus P2B-S Mainboard and since a week I have a Maxtor 60 GB
> 
> that's piix-based, is it not?

It is BX-chipset, not exactly sure if this is what is called piix. But I
think so and found this in /proc/ide/pix

                                Intel PIIX4 Ultra 33 Chipset.
--------------- Primary Channel ---------------- Secondary Channel
-------------                 enabled                          enabled
--------------- drive0 --------- drive1 -------- drive0 ---------- drive1
------DMA enabled:    yes              no              yes               no 
UDMA enabled:   yes              no              yes               no 
UDMA enabled:   2                X               2                 X
UDMA
DMA
PIO

> > >  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
> 
> should be more like 30 MB/s, so you must not have the piix controller
> in your kernel.

I was expecting something above 20MB/s, and thus I wondered about the low
values. Well and when I tried first time, my computer was not really idle
and thus I only got something around 3.5 MB/s.
 

OK, what do I need to compile into my kernel to have the chipset configured
fine? Or do I need any boot-parameters?

Here is an excerpt from my kernel-config:

CONFIG_IDE=y
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_IDEDISK_MULTI_MODE=y
CONFIG_BLK_DEV_IDECD=m
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_IDEPCI_SHARE_IRQ=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
CONFIG_BLK_DEV_ADMA=y
CONFIG_BLK_DEV_OFFBOARD=y
CONFIG_IDEDMA_PCI_AUTO=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_BLK_DEV_PIIX=y
CONFIG_PIIX_TUNING=y
CONFIG_IDE_CHIPSETS=y
CONFIG_IDEDMA_AUTO=y
CONFIG_BLK_DEV_IDE_MODES=y

everything skipped is just not set.

Thanks, thomas
-- 
thomas@deselaers.de «<>» JabberID on request «<>» GPG/PGP key on request
  «< Unless stated otherwise everything I write is just my opinion >»

Re: IDE Harddrive Performance

[Mike Dresser]

> I dont really know, I dont think its possible to get higher then that from a
> 5400 RPM disk. Heres mine,
>
> FUJITSU 40.9GB MPG3409AT 5400 RPM
>
> /dev/hda:
> Timing buffered disk reads:  64 MB in  7.96 seconds =  8.04 MB/sec

Maxtor 60.0GB 96147U8 5400 RPM (Diamond Max 60, running in ultra/33 mode)

/dev/hdd:
 Timing buffered disk reads:  64 MB in  3.45 seconds = 18.55 MB/sec

It's been a long time since i looked at the machine, so I don't remember
the mb type, it's using the PIIX4 controller, celeron 500.  The bios
doesn't support over 32 gb, so it's set to none in the bios, and Andre's
patches help Linux pick it up and run optimized.  Stock 2.2.19/20 kernels
run at about half this speed.

mike

Re: IDE Harddrive Performance

[Oleg Artamonov]

Thomas Deselaers написал:
> /dev/hdc:
>  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
>
> What would be a value I can expect from my hardware?

  I have a Maxtor D541X (20Gbytes, 5400rpm), and its results are about 
37MB/sec... Motherboard is Epox 8KTA3 with VIA 686B southbridge (HDD runs in 
UDMA5 mode).

> And what might result in higher speeds?

  Did you enable UltraDMA2 (P2B-S supports only UDMA2, not UDMA4 nor UDMA5)? 
32-bit transfer? What 'hdparm /dev/hda' and 'hdpram -i /dev/hda' says?

Re: IDE Harddrive Performance

[Thomas Deselaers]

On Wed, Dec 19, 2001 at 06:55:38PM +0300, Oleg Artamonov wrote:
> Thomas Deselaers ???????:
> > /dev/hdc:
> >  Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec
> >
> > What would be a value I can expect from my hardware?
> 
>   I have a Maxtor D541X (20Gbytes, 5400rpm), and its results are about 
> 37MB/sec... Motherboard is Epox 8KTA3 with VIA 686B southbridge (HDD runs in 
> UDMA5 mode).
> 
> > And what might result in higher speeds?
> 
>   Did you enable UltraDMA2 (P2B-S supports only UDMA2, not UDMA4 nor UDMA5)? 
> 32-bit transfer? What 'hdparm /dev/hda' and 'hdpram -i /dev/hda' says?

Here are the results, and maybe I should have told, that I am using 2.4.16
without any further patches applied. 

It seems that I have udma2 enabled.

leukertje:/home/thomasd# hdparm /dev/hdc

/dev/hdc:
 multcount    = 16 (on)
 I/O support  =  0 (default 16-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 7297/255/63, sectors = 117231408, start = 0
 busstate     =  1 (on)
leukertje:/home/thomasd# hdparm -i /dev/hdc

/dev/hdc:

 Model=MAXTOR 4K060H3, FwRev=A08.1500, SerialNo=373120024159
 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
 RawCHS=16383/16/63, TrkSize=32256, SectSize=21298, ECCbytes=4
 BuffType=DualPortCache, BuffSize=2000kB, MaxMultSect=16, MultSect=16
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=117231408
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes: pio0 pio1 pio2 pio3 pio4 
 DMA modes: mdma0 mdma1 mdma2 udma0 udma1 *udma2 
 AdvancedPM=no WriteCache=enabled
 Drive Supports : ATA/ATAPI-5 T13 1321D revision 1 : ATA-1 ATA-2 ATA-3 ATA-4
ATA-5 


thomas

-- 
thomas@deselaers.de «<>» JabberID on request «<>» GPG/PGP key on request
  «< Unless stated otherwise everything I write is just my opinion >»

Re: IDE Harddrive Performance

[Oleg Artamonov]

Thomas Deselaers написал:

> Here are the results, and maybe I should have told, that I am using 2.4.16
> without any further patches applied.

  Yes, i'm using 2.4.8 form Mandrake distribution.

> leukertje:/home/thomasd# hdparm /dev/hdc
>
> /dev/hdc:
>  multcount    = 16 (on)
>  I/O support  =  0 (default 16-bit)
		^^^
  'hdparm -c1 /dev/hda' to enable 32-bit access - it will be much faster.

Re: IDE Harddrive Performance

[Thomas Deselaers]

On Wed, Dec 19, 2001 at 07:15:34PM +0300, Oleg Artamonov wrote:
> > leukertje:/home/thomasd# hdparm /dev/hdc
> >
> > /dev/hdc:
> >  multcount    = 16 (on)
> >  I/O support  =  0 (default 16-bit)
> 		^^^
>   'hdparm -c1 /dev/hda' to enable 32-bit access - it will be much faster.

did this, but with no success, drive is not becoming faster:
leukertje:/home/thomasd# hdparm -c1 /dev/hdc

/dev/hdc:
 setting 32-bit I/O support flag to 1
 I/O support  =  1 (32-bit)
 leukertje:/home/thomasd# hdparm -t /dev/hdc

dev/hdc:
 Timing buffered disk reads:  64 MB in  5.62 seconds = 11.39 MB/sec
   
Thanks,
thomas
-- 
thomas@deselaers.de «<>» JabberID on request «<>» GPG/PGP key on request
  «< Unless stated otherwise everything I write is just my opinion >»

Re: IDE Harddrive Performance

[Holger Lubitz]

> /dev/hdc:
>  Timing buffered disk reads:  64 MB in  5.63 seconds =3D 11.37 MB/sec

Thats low. I get 25.5 MB/sec from my old Maxtor 96147U8 on a P2L97-DS
(this was Maxtors first 60 gig drive - 8 heads, 15 gig per platter -
yours has 3 heads with 40 gigs per platter. the higher recording density
should make it even faster than mine).

But: This is still more than my IBM DTLA does on that board - which in
itself is perfectly capable of 36 MB/sec when driven with UDMA 100, but
drops to approx. 22 MB/sec when limited to UDMA33. So your drive might
have a similar issue. Its nominal transfer rate is higher than UDMA 33
can deliver, and it wastes bandwidth in adapting to the slower bus.

Holger

Re: IDE Harddrive Performance

[Adam Keys]

On December 19, 2001 09:32, Thomas Deselaers wrote:
> I do have an Asus P2B-S Mainboard and since a week I have a Maxtor 60 GB
> 5400 rpm Harddrive (MAXTOR 4K060H3).

I am experiencing extremely shoddy performance with a similar disk/chipset 
combination.  I am on an Asus P2B-DS board (440BX) with the following disk 
setup.

hda: Maxtor 93073U6, ATA DISK drive
hdc: Maxtor 90840D6, ATA DISK drive
hdd: Maxtor 90840D6, ATA DISK drive

I am barely experiencing the same performance you are, after enabling all the 
stuff that's been discussed in this thread.    I'd like to determine if this 
disk is dying or if there is something seriously wrong with my kernel.  I did 
not seem to experience anything until I went to 2.4.16.  hdd seems to be 
headed for the heap, as all tests on it fail.  I am stumped; I've read the 
manual and the threads and still haven't extracted proper performance.  All 
help is appreciated.

Thanks,
AKK

Following is the information on the other disks and the hdparm test results:

# hdparm -t /dev/hda

/dev/hda:
 Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec

# hdparm /dev/hda ; hdparm -i /dev/hda

/dev/hda:
 multcount    = 16 (on)
 I/O support  =  1 (32-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 3736/255/63, sectors = 60030432, start = 0
 busstate     =  1 (on)

/dev/hda:

 Model=Maxtor 93073U6, FwRev=FA500S60, SerialNo=G607E80C
 Config={ Fixed }
 RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
 BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=60030432
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes: pio0 pio1 pio2 pio3 pio4
 DMA modes: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4
 AdvancedPM=yes: disabled (255) WriteCache=enabled
 Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3 ATA-4 
ATA-5


/dev/hdc:
 Timing buffered disk reads:  64 MB in  5.63 seconds = 11.37 MB/sec

/dev/hdc:
 multcount    = 16 (on)
 I/O support  =  1 (32-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 nowerr       =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 1021/255/63, sectors = 16406208, start = 0
 busstate     =  1 (on)

/dev/hdc:

 Model=Maxtor 90840D6, FwRev=WAS82739, SerialNo=K606M2WA
 Config={ Fixed }
 RawCHS=16276/16/63, TrkSize=0, SectSize=0, ECCbytes=29
 BuffType=DualPortCache, BuffSize=256kB, MaxMultSect=16, MultSect=16
 CurCHS=16276/16/63, CurSects=16406208, LBA=yes, LBAsects=16406208
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes: pio0 pio1 pio2 pio3 pio4
 DMA modes: mdma0 mdma1 mdma2 udma0 udma1 *udma2
 AdvancedPM=no WriteCache=enabled
 Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3 ATA-4
-- 
akk~

Re: IDE Harddrive Performance

[Alan Cox]

> # hdparm -t /dev/hda
> 
> /dev/hda:
>  Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec

Do you get sane numbers if you use 2.4.9 for the hdparm test ?

Re: IDE Harddrive Performance

[Adam Keys]

On December 22, 2001 02:56, Alan Cox wrote:
> > # hdparm -t /dev/hda
> >
> > /dev/hda:
> >  Timing buffered disk reads:  64 MB in 86.98 seconds =753.46 kB/sec
>
> Do you get sane numbers if you use 2.4.9 for the hdparm test ?

Just tried 2.4.9 and 2.4.10, both with ext3 patches.  Both had the same 
abysmal numbers (791.48 KB/sec for 2.4.9).

Any other combinations I should try?
-- 
akk~

Re: IDE Harddrive Performance

[Thomas Deselaers]

On Sat, Dec 22, 2001 at 03:09:05PM -0200, vda wrote:
> I think it makes sense for anyone who suspect bad hdd performance due to 
> Linux to perform tests under Windows. My hdd does not work any better in Win 
> so I conclude it may be drive problem, not OS.

What is a program to run tests under Windows which gives comparable results?

greeting,
thomas

-- 
thomas@deselaers.de «<>» JabberID on request «<>» GPG/PGP key on request
  «< Unless stated otherwise everything I write is just my opinion >»