[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Pci to Pci transfers
- To: Mailing List Recipients <pci-sig-request@znyx.com>
- Subject: RE: Pci to Pci transfers
- From: Jim Ulrich <jimu@pathlight.com>
- Date: Wed, 11 Sep 1996 10:46:13 -0400
- Resent-Date: Wed, 11 Sep 1996 10:46:13 -0400
- Resent-From: pci-sig-request@znyx.com
- Resent-Message-Id: <"yFTug.0.Hw4.i3jDo"@dart>
- Resent-Sender: pci-sig-request@znyx.com
----------
From: Herve Suquet[SMTP:hsu@ocegr.fr]
Sent: Wednesday, September 11, 1996 3:07 AM
To: Mailing List Recipients
Subject: Pci to Pci transfers
Hello PCI expert,
....
I received two answers. One saying that it is obvious to do and without risks
and one saying that a lot of problem have to be faced-up making this
almost impossible to run on various PC platforms.
Between these two answers, where is the good one ?
With best regards,
Herve
#############################################
# Herve SUQUET #
# ES-GA - OCE Creteil #
# 1, rue Jean Lemoine - B.P.113 #
# 94003 Creteil Cedex - FRANCE #
# Tel (standard) : (33) (1) 48 98 80 00 #
# Tel (direct) : (33) (1) 48 98 80 06 #
# Fax : (33) (1) 48 98 54 50 #
# Email : suquet@ocegr.fr #
#############################################
Herve,
At Pathlight we do a lot with direct transfer from one PCI device
(ours) to another without host intervention. The caveats we have are:
1] The host should be 2.1 compliant if any of these devices
is behind a P2P bridge (as many of ours are).
2] If the PCI devices are in a Power-MAC (or clone) then use
of the MWI MRM MRL commands are needed to get ANY sort
of reasonable performance.
In a product that I just finished bringing up we have several PCI
devices behind a DEC bridge. A majority of the traffic is between
these devices, with virtually no transfers to/from the host. This
eliminates host arbitration and bandwidth bottle-necks.
Jim Ulrich
jimu@pathlight.com
My brain cells are my own and do not reflect those of my company...
è
û