public inbox for linuxppc-dev@ozlabs.org 
 help / color / mirror / Atom feed
From: Gerhard Jaeger <g.jaeger@sysgo•com>
To: linuxppc-embedded@ozlabs•org
Cc: "Martin, Tim" <tim.martin@viasat•com>
Subject: Re: PPC Linux support for Tundra TSI148
Date: Thu, 27 Apr 2006 09:13:10 +0200	[thread overview]
Message-ID: <200604270913.10778.g.jaeger@sysgo.com> (raw)
In-Reply-To: <821B2170E9E7F04FA38DF7EC21DE4871049711CD@VCAEXCH01.hq.corp.viasat.com>

See my comments inline:

On Wednesday 26 April 2006 18:24, Martin, Tim wrote:
> Gerhard,
> 
> Thanks for the information.  More questions...
> 
> What type of transfers were you doing? (e.g. A32/D64? SST320 or SST267?)
A32/D64 and 2eSST @ SST320

> Are these transfer from userspace data, from kernelspace data, or from
> the Tundra's pattern buffer?

They are from userspace.

> What was your PCI bus speed & width? 
???

> Were you using inbound/outbound windows or Tundra's DMA controller?
DMA - via /dev/vme_dma0 interface

> What was the Tundra chipset configuration for the 168 MBps?
i.e. Buffersize 64K, 4K alignment, 2K PCI Fifo and 1K VME Fifo.
Using the Tundra driver and its interface structure, I used a 
request level of 2, fairmode is set to normal, timeon timeout timer
is set to 7 and timeoff is set to no waiting (0).

> 
> Thanks,
> Tim
> 
> > -----Original Message-----
> > From: linuxppc-embedded-bounces+tmartin=viasat.com@ozlabs•org
> > [mailto:linuxppc-embedded-bounces+tmartin=viasat.com@ozlabs•org] On
> Behalf
> > Of Gerhard Jaeger
> > Sent: Wednesday, April 26, 2006 12:26 AM
> > To: linuxppc-embedded@ozlabs•org
> > Subject: Re: PPC Linux support for Tundra TSI148
> > 
> > On Wednesday 26 April 2006 01:47, Martin, Tim wrote:
> > > Does anyone out there have any real world measured performance of a
> > > Linux PowerPC (kernel module + user space application) doing 2eSST
> VME
> > > transfers with the Tundra TSI148 chipset?
> > >
> > > Tundra has a Linux driver available for the Motorola MVME6100 , but
> told
> > > me they don't have any performance data available.  I'm looking for
> > > sustained throughput rates, not the peak burst rates (e.g. 320 MBps,
> 267
> > > MBps).
> > >
> > 
> > Hi,
> > 
> > I've done here some tests between two MVME6100, while updating the
> 6100
> > BSP
> > for our embedded Linux distro. Depending on the buffersize, alignment,
> VME
> > and
> > PCI FIFO settings we have (without further optimizations) throughput
> rates
> > ranging
> > from 100MBps up to 168MBps. I think some tuning could still be done.
> > 
> >HTH
> > Gerhard

-- 
Gerhard Jaeger <gjaeger@sysgo•com>            
SYSGO AG                      Embedded and Real-Time Software
www.sysgo.com | www.elinos.com | www.pikeos.com | www.osek.de 

  reply	other threads:[~2006-04-27  7:13 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-04-26 16:24 PPC Linux support for Tundra TSI148 Martin, Tim
2006-04-27  7:13 ` Gerhard Jaeger [this message]
  -- strict thread matches above, loose matches on Subject: below --
2006-04-25 23:47 Martin, Tim
2006-04-26  7:26 ` Gerhard Jaeger

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200604270913.10778.g.jaeger@sysgo.com \
    --to=g.jaeger@sysgo$(echo .)com \
    --cc=linuxppc-embedded@ozlabs$(echo .)org \
    --cc=tim.martin@viasat$(echo .)com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox