SCSI Performance in vxWorks

SCSI Performance in vxWorks

Post by Jared Maguir » Wed, 23 Jan 2002 15:58:18



Has anyone measured SCSI performnace greater than 10MB/s under vxWorks?
What card/driver?

We're using an LSI Logic 53C1010 controller, and couple of Seagate
Cheetahs. We've also tried it on a sym895, and an adaptec 7880, and we
can't seem to get a transfer to happen at greater than 10MB/s, sustained.
What gives?

-J

 
 
 

SCSI Performance in vxWorks

Post by David Clar » Wed, 23 Jan 2002 22:01:55



> Has anyone measured SCSI performnace greater than 10MB/s under vxWorks?
> What card/driver?

The controller/drive is only part of the equation - the other part is
going to be the type of connection/interface.  If the board or drive only
supports Fast SCSI, then 10MB/s is going to be fastest you can go.

If you can do Fast Wide (16 bit), then you can get up to 20MB/s - and Wide
Ultra will get you up to a max. possible of 40MB/s.

Using an Ultra Wide Western Digital drive with a DY 4 SVME/DMV-179 SBC
configured for 16-bit SCSI, we've gotten > 30MB/s (read/write).

--
-------------------------------  /'''  --------------------------------
David Clarke                     c-00

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

 
 
 

SCSI Performance in vxWorks

Post by Tim Sha » Thu, 24 Jan 2002 07:45:38


We've had up to about 17MBytes/sec on a Synergy VGM5 (the driver had been
modified by Synergy), and about the same on a MVME2307 with SYM875 (we had to
get hold of the source code and modify it to get that performance). I've
never used the 895, but if it is using the same driver as the wr$ driver for
the mv2307 there's a lot of tweaking possible to get big performance
improvements (both increasing data transfer rate and reducing pci usage).

HTH
Tim


> Has anyone measured SCSI performnace greater than 10MB/s under vxWorks?
> What card/driver?

> We're using an LSI Logic 53C1010 controller, and couple of Seagate
> Cheetahs. We've also tried it on a sym895, and an adaptec 7880, and we
> can't seem to get a transfer to happen at greater than 10MB/s, sustained.
> What gives?

> -J

 
 
 

SCSI Performance in vxWorks

Post by Jared Maguir » Fri, 25 Jan 2002 09:13:04




> We've had up to about 17MBytes/sec on a Synergy VGM5 (the driver had
> been modified by Synergy), and about the same on a MVME2307 with SYM875
> (we had to get hold of the source code and modify it to get that
> performance). I've never used the 895, but if it is using the same
> driver as the wr$ driver for the mv2307 there's a lot of tweaking
> possible to get big performance improvements (both increasing data
> transfer rate and reducing pci usage).

<nod>. Just now I timed scsi performance under both Win2k and vxWorks on
the same hardware, and Win2k performed it's 4MB write in about 100ms,
which is pretty decent. VxWorks performed the same write in 400ms. I dove
in with windview to try to see what's going on, and found that the SCSI
driver's interrupt handler gets called a few times at each of THREE points
during a write-operation. These points are at t=0ms, t=100ms and t=400ms.
The t=100ms point is interesting, as that's EXACTLY how long windows takes
to write the data. I'm starting to suspect that the scsi driver is
actually writing the data at a reasonable speed, but then is doing
SOMETHING else before it returns successfully. *shrug*

That's a bummer that we might need to get the code for the driver and
tweak it. Oh, well.

Thanks for the reality-check,

-Jared

 
 
 

SCSI Performance in vxWorks

Post by Leonid Rosenboi » Mon, 28 Jan 2002 16:18:42


One possible guess is that SOMETHING could be verifying data that just has
been written,
but I doubt that the driver does anything useful during it's extra time.
Wind River has been allways kinda shaky in this area, if you read this group
for a long time you'd notice that device driver is certainly not one of
WRS's strong
areas, and knowledge is much weaker then other driver areas.

In other words, if you need a good driver, you'd be better off paying
someone
to develop o optimize it for you, e.g. one of former WRS developers now
freelancing.

- Leonid




> > We've had up to about 17MBytes/sec on a Synergy VGM5 (the driver had
> > been modified by Synergy), and about the same on a MVME2307 with SYM875
> > (we had to get hold of the source code and modify it to get that
> > performance). I've never used the 895, but if it is using the same
> > driver as the wr$ driver for the mv2307 there's a lot of tweaking
> > possible to get big performance improvements (both increasing data
> > transfer rate and reducing pci usage).

> <nod>. Just now I timed scsi performance under both Win2k and vxWorks on
> the same hardware, and Win2k performed it's 4MB write in about 100ms,
> which is pretty decent. VxWorks performed the same write in 400ms. I dove
> in with windview to try to see what's going on, and found that the SCSI
> driver's interrupt handler gets called a few times at each of THREE points
> during a write-operation. These points are at t=0ms, t=100ms and t=400ms.
> The t=100ms point is interesting, as that's EXACTLY how long windows takes
> to write the data. I'm starting to suspect that the scsi driver is
> actually writing the data at a reasonable speed, but then is doing
> SOMETHING else before it returns successfully. *shrug*

> That's a bummer that we might need to get the code for the driver and
> tweak it. Oh, well.

> Thanks for the reality-check,

> -Jared

 
 
 

1. SCSI->SCSI vs. SCSI->ESDI disk performance?

I have run into an apparently unresolvable question.
Is there a performance hit in having an ESDI disk
hooked to a SCSI bus through a translator (Emulex MD21)
versus a SCSI disk hanging directly on the bus?

ESDI disk performance is supposed to be higher than
that of SCSI disks.  Does this mean that the two
configurations are roughly equivalent in performance
after taking into account the MD21?

The questions burn on and on.

Thanks.  Tom.
--
Tom Poage, UCDMC Clinical Engineering, Sacto., CA

2. IIR's

3. MVME167 SCSI performance under Vxworks

4. New MSN Explorer v6.10.0016.1625 - Released

5. VxWorks SCSI Performance

6. Multiplexer Software Control

7. SCSI disk sparing algorithms (was: Disk performance issues, was IDE vs SCSI-2 using iozone (long))

8. Functions help

9. Adaptec SCSI 1522 SCSI host adapterm NEC SCSI external 6x CD_ROM

10. SCSI,SCSI II, WIDE SCSI

11. scsi vs. wide scsi vs. fast and wide scsi?

12. HP7400C: Best Performance - USB or SCSI?