🚀 go-pugleaf

In article <1992Dec17.153141.3926@urbana.mcd.mot.com> dfields@urbana.mcd.mot.com writes:
>In article <1992Dec16.211712.13142@twisto.eng.hou.compaq.com>, simonich@croatia.eng.hou.compaq.com (Chris Simonich) writes:
>> In article <1gntdfINNu7@cbl.umd.edu> mike@cbl.umd.edu (Michael Santangelo) writes:
>>>Can EISA, MicroChannel, or the new VESA LocalBus on PC's allow
>>multiple controllers plugged into them (say two smart bus mastering
>>>SCSI disk controllers on one of these buses) xfer data simultaneously
>>>to main memory (DMA)?
>>>
>>Generally, shared bus architectures are shared by time-division multiplexing.
>>In short, only one device may use the "bus" at a time, so the answer to your
>>question is No.
>
>I read waht he said and wasn't so sure from the example he gave that the
>question he asked is the one he wanted to ask.  To controllers can be doing
>DMA trnasfers at the same time byspliting the transfers into smaller chunks
>and doing the "time-division multiplexing" of the bus via an arbitraition
>scheme.
>
>>There was a post about a new I/O architecture from Compaq.  This architecture
>>does allow multiple controllers to perform simultaneously.
>
>If the MIOC in Compaq's proposal is implemented with a multi-port interface
>to main memory then it will.  You would loose some on cost of course.
>--
>Dave Fields	Motorola Computer Group		dfields@urbana.mcd.mot.com
>SVR4: The Winchester Mystery Kernel		uiucuxc!udc!dfields

A multi-port interface to memory is not required.  If two devices on an IOC
are performing DMA simultaneously, all that is required is that the IOC
send the data upstream at a rate equal to that of the total of the devices.
There is nothing preventing the upstream channel on the IOC from running
at 100 MHz.  Of course, another way to solve the problem would be to use
additional buffering in the IOC, for example.  The point is that from the
devices point of view it doesn't have to share its channel with anybody at
any time.
--
======================================================
Christopher Simonich        simonich@twisto.compaq.com
Compaq Computer Corp.       [713] 374-1898
======================================================