@Amike

Nice picture, but would it be possible for you to just copy and paste the output of scsispeed here? Much better for everyone to read them. Yes? Thanks!

OldFart

_________________
More then three levels of indigestion and you're scroomed!

This is off the same µA1 as above, I've not touched the IDE settings, but have installed the new IDE patch from Hyperion.


6.RAM Disk:Scsispeed4> scsispeed DRIVE a1ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 12408012 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 12885196 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 13066240 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 13107200 bytes/sec

_________________
Rage quited 29th May 2011

@Stephen_Robinson

Thanks! Figures updated.

EDIT: Figures are still very impressive in the negative sense of the word!

OldFart

Last edited by OldFart on 10-Mar-2009 at 09:01 PM.

_________________
More then three levels of indigestion and you're scroomed!

@OldFart

Quote:

OldFart wrote: @Stephen_Robinson DMA can be set with this variable in UBoot cli: a1ide_xfer. First, however, you should check the value set with this command from your Amiga-shell: nvgetvar a1ide_xfer OldFart

Just tried that, and it comes back
0 0

What should I do now??

Just to prove I'm not useless, I've checked it out the Interwebs

http://www.titan.co.nz/amigaak/AA020816.htm

nd that sounds like the correct setting too.

Last edited by Stephen_Robinson on 14-Mar-2009 at 04:02 PM.

_________________
Rage quited 29th May 2011

@OldFart

Since an encounter of the third kind with UBoot Prefs in September, I have been running my system in best PIO mode. Last night I finally took the last step to go back to UDMA 5 (100 MB/s) mode for my hard drive.

MicroA1-C with onboard VIA IDE controller.

setenv a1ide_maxbus 2
setenv a1ide_timeout 5
setenv a1ide_conf 1020
setenv a1ide_xfer F000
saveenv

also had to change boot sequence because I entered U-boot with wrong key.


2.AmigaOS4:> idetool -u a1ide.device 0
-------------------Unit 0 information------------------------
Flags : $00000187 - present, LBA48 used, supports DMA, ATA, supports cache flush, interrupts used,
Xfer mode : best pio 12 (PIO 4, 16 MB/s) / best dma 69 (UDMA 5, 100 MB/s) / current 69 (UDMA 5, 100 MB/s)
Total blocks : 488397168
Blocksize : 512
Power mode : 2 / idle (ready for operation)
IO1 / IO2 / BMCR @ : $1F0 / $3F6 / $CC00



Here are the new results for Scsispeed4 test:

2.AmigaOS4:> ram:scsispeed4/scsispeed DRIVE a1ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 28934144 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 34888089 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 38966067 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 41114009 bytes/sec
2.AmigaOS4:>

---
redfox

@redfox

Right, before:-

6.RAM Disk:> idetool -u a1ide.device 0
-------------------Unit 0 information------------------------
Flags : $00000185 - present, supports DMA, ATA, supports cache flush, interrupts used,
Xfer mode : best pio 12 (PIO 4, 16 MB/s) / best dma 69 (UDMA 5, 100 MB/s) / current 12 (PIO 4, 16 MB/s)
Total blocks : 156301488
Blocksize : 512
Power mode : 2 / idle (ready for operation)
IO1 / IO2 / BMCR @ : $1F0 / $3F6 / $CC00

6.RAM Disk:Scsispeed4> scsispeed DRIVE a1ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 12408012 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 12885196 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 13066240 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 13107200 bytes/sec

I'll be back in a mo..

And after...

idetool -u a1ide.device 0
-------------------Unit 0 information------------------------
Flags : $00000185 - present, supports DMA, ATA, supports cache flush, interrupts used,
Xfer mode : best pio 12 (PIO 4, 16 MB/s) / best dma 69 (UDMA 5, 100 MB/s) / current 69 (UDMA 5, 100 MB/s)
Total blocks : 156301488
Blocksize : 512
Power mode : 2 / idle (ready for operation)
IO1 / IO2 / BMCR @ : $1F0 / $3F6 / $CC00


6.RAM Disk:> scsispeed drive a1ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 20546764 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 24041062 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 24682496 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 26315980 bytes/sec

Last edited by Stephen_Robinson on 14-Mar-2009 at 10:22 PM.
Last edited by Stephen_Robinson on 14-Mar-2009 at 10:21 PM.

_________________
Rage quited 29th May 2011

@Stephen_Robinson

Quote:

Just tried that, and it comes back 0 0

One thing: there should be NO blanks in the setting!
Mine (µA1) returns F0E0, meaning HD on first fort is specified for UDMA 5, the second port is no in use, the third port (CD-RW) is specified for UDMA 4 and the fourth port is not in use.

In you case I would set to something simmilar.

What does this command return?
nvgetvar a1ide_conf

New Shell process 10
10.OS_4.1:> nvgetvar a1ide_xfer
F0E0
10.OS_4.1:> nvgetvar a1ide_conf
1020
10.OS_4.1:> 

_________________
More then three levels of indigestion and you're scroomed!

@OldFart
Sam440/OS4.1 with Samsung Spinpoint 160GB SATA:

scsispeed DRIVE sii3114ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: sii3114ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 30064640 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 49928601 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 70103859 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 78938112 bytes/sec


---
whatelse

@redfox

Congratulations! A speed increase of about a magnitude of 3! (Actually 2, for 1 was already present and it is now 3, so...)
You won't regret your decisions, now will you? Btw, do you 'feel' the speed increase?

Now where's that darn dancing banana?

OldFart

_________________
More then three levels of indigestion and you're scroomed!

@Stephen_Robinson

About twice as fast as it used to be, but still not *that* impressive. Could you go one step higher and see what the effects are?

OldFart

_________________
More then three levels of indigestion and you're scroomed!

@OldFart

[spinal Tap]Can you go one higher? [/spinal tap]

"6.RAM Disk:> nvgetvar a1ide_conf
1020
6.RAM Disk:> nvgetvar a1ide_xfer
F0E0
"

According to that website I linked above, F is the highest it goes:-

1ide_xfer
(not set)
4 chars - (primary master,primary slave,secondary master,secondary slave)
  where 0 = Automatic
a = PIO 0
b = PIO 1
c = PIO 2
d = PIO 3
e = PIO 4
A = UDMA 0
B = UDMA 1
C = UDMA 2
D = UDMA 3
E = UDMA 4
F = UDMA 5
(let si680ide choose best PIO mode)
(3 MB/s, modeid 8)
(5 MB/s, modeid 9)
(8 MB/s, modeid 10)
(11 MB/s, modeid 11)
(16 MB/s, modeid 12)
(16 MB/s, modeid 64)
(25 MB/s, modeid 65)
(33 MB/s, modeid 66)
(44 MB/s, modeid 67)
(66 MB/s, modeid 68)
(100 MB/s, modeid 69)
Onboard VIA686B ONLY - specifies the transfer mode that a1ide.device will use for each device.
If not set, a1ide.device will use the best PIO mode supported by the drive.
If you specify an unsupported mode, it will use the best mode the drive claims to support.

Last edited by Stephen_Robinson on 14-Mar-2009 at 10:42 PM.

_________________
Rage quited 29th May 2011

@OldFart

Thanks. Yes, it is significantly faster. I want to let the system "soak" for a few more days, before I let out the dancing bananas. (I've had a few bananas go up in flames in the past after I spoke too soon. )

---
redfox

@Stephen_Robinson

I am not an expert by any means, but maybe you should try a slightly lower UDMA setting, such as E for UDMA 4 (66 MB/s, modeid 68) to see if that makes any difference.

Even though it reports best dma 69 (UDMA 5, 100 MB/s), maybe the hard drive requires more support than is actually available from our current driver.

---
redfox

Last edited by redfox on 15-Mar-2009 at 02:45 AM.

@redfox

Just set it to E0E0

and 6.RAM Disk:> scsispeed DRIVE a1ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 29461299 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 35215769 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 38436864 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 41022259 bytes/sec


That'll do for me.

_________________
Rage quited 29th May 2011

@Stephen_Robinson

Quote:

That'll do for me.

Must've been an eye goggling experience? Avery near-exact tripling of diskwise performance. Congratulations!


OldFart

_________________
More then three levels of indigestion and you're scroomed!

Here's the results of my new Sam440ep@600MHz, using a OCZ SSD Series 2 30GB disk:

4.System4.1:Utilities> scsispeed DRIVE sii3114ide.device:0 FAST BUF1=8192 BUF2=16384 BUF3=32768 BUF4=65536
MKSoft ScsiSpeed 4.2 Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020 AmigaOS Version: 53.5
Device: sii3114ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 28816588 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 55750656 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 69070028 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI: 78702182 bytes/sec

@OldFart

** BUMP **

Any newbees around or other people that have currently nothing more to do then 'picking their nose'?

As it currently stands, SAM440 n combo with Seagate is the absolute winner. However, I'm very curious for the performance of a SAM440 and a WD Raptor. Somebody fiddling with this combo, perhaps?

OldFart

_________________
More then three levels of indigestion and you're scroomed!

AmigaOne-XE G3 800MHz
AmigaOS 4.1 (Kickstart 53.5 Workbench 53.1)
HDD: Samsung SP0411N 40 GB on a1ide.device
DMA 69 (UDMA 5, 100 MB/s) and IRQ enabled

MKSoft ScsiSpeed 4.2  Copyright © 1989-92 MKSoft Development
AmigaOS4 port by Stéphane Guillard
------------------------------------------------------------
CPU: 68020  AmigaOS Version: 53.5
Device: a1ide.device:0


Testing with a 8192 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI:  25926041 bytes/sec

Testing with a 16384 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI:  31730073 bytes/sec

Testing with a 32768 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI:  35894067 bytes/sec

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.
Read from SCSI:  38384435 bytes/sec

Last edited by amigafan on 17-Apr-2009 at 10:09 PM.
Last edited by amigafan on 17-Apr-2009 at 09:59 PM.

@amigafan

Thank you!

(And bumpeteebump goes the thread... )

OldFart

_________________
More then three levels of indigestion and you're scroomed!

@Spirantho

Quote:

Testing with a 65536 byte, MEMF_FAST, LONG-aligned buffer.Read from SCSI: 32974438 bytes/sec

So now we have another WD drive acting slow, in SATA, which should double the throughput.