[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Burst reads required?
- To: Mailing List Recipients <pci-sig-request@znyx.com>
- Subject: RE: Burst reads required?
- From: Alexei Predtechenski <Alexei.Predechenski@amd.com>
- Date: Tue, 3 Sep 1996 09:03:06 -0500
- Encoding: 40 TEXT
- Resent-Date: Tue, 3 Sep 1996 09:03:06 -0500
- Resent-From: pci-sig-request@znyx.com
- Resent-Message-Id: <"W0iJ7.0.iq1.dj3Bo"@dart>
- Resent-Sender: pci-sig-request@znyx.com
>----------
>From: weathert@arl.wustl.edu[SMTP:weathert@arl.wustl.edu]
>Sent: Wednesday, August 28, 1996 2:01 PM
>To: Mailing List Recipients
>Subject: Burst reads required?
>
>My name is William Eatherton and I am devloping a target PCI synthesis
>core.
>[...some stuff deleted...]
> If DEC Alphas, PowerPCs, Intel PCs, and Intel clones all generate
>non-burst reads than I am thinking of not implementing the option of
>burst reads at all.
I do not think it is a wise decision. The CPU is not the only master on
the PCI bus. If your target has a memory-mapped buffer, it can be
directly accessed by another PCI bus-master (say, disk controller) in a
perfect burst read mode, and your design will be very "suboptimal" in
terms of bus utilization.
- Alex
_________________________ _______
Alex Predtechenski \____ | Advanced
PP Application Engineer, /| | | Micro
CPG Systems Engineering and Validation | |___| | Devices
(512)602-3567 |____/ \|
5900 E. Ben White Blvd, m/s 571 Austin, TX 78741
============================================================
Disclaimer: The views expressed are mine, not necessarily
those of my employer.
All trademarks are acknowledged, etc
>
>
>
· ü é