Browse Prior Art Database

Optimized Cycle Times on a Multi-Drop Main Storage Data Bus

IP.com Disclosure Number: IPCOM000106382D
Original Publication Date: 1993-Nov-01
Included in the Prior Art Database: 2005-Mar-21
Document File: 2 page(s) / 58K

Publishing Venue

IBM

Related People

Diffield, NJ: AUTHOR [+5]

Abstract

A method to optimize cycle times between devices on a multi-drop bus is described. Instead of running the whole bus at the cycle time dictated by the slowest device to device transfer on the bus, the bus is optimized to run at the cycle time dictated by device to device transfer rate where the traffic is heaviest, with less frequent transfers running at the slower cycle time dictated by the slowest device to device transfer rate on the bus.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 56% of the total text.

Optimized Cycle Times on a Multi-Drop Main Storage Data Bus

      A method to optimize cycle times between devices on a
multi-drop bus is described.  Instead of running the whole bus at the
cycle time dictated by the slowest device to device transfer on the
bus, the bus is optimized to run at the cycle time dictated by device
to device transfer rate where the traffic is heaviest, with less
frequent transfers running at the slower cycle time dictated by the
slowest device to device transfer rate on the bus.

      When implementing a bus with multiple drops (more than two
devices) the delay required to pass information between pairs of
devices will vary.  When a single cycle time is selected for
transfers between all devices on the bus, the cycle time must be set
to allow for the longest delay between any two devices on the bus.
In the diagram (Fig. 1), the distance between devices A, B, and C is
quite short compared to the distance from them to device D.  This can
happen for example, if device A, B, and C are on the same multi-chip
module carrier, while device D is on another card.  This bus
configuration requires longer to transfer data between the devices A,
B, and C, versus the delay for communication from these devices to
device D, due to having to wait for the reflection from the far end
of the bus to return.  In the example, the volume of traffic from
devices A, B, and C to device D is much greater than the traffic
between devices A, B, and C.  Operating...