Architectural tricks to maximize memory bandwidth

MessageThis Webinar is over
Date Feb 17, 2016
Time 11:00 am PST
Cost Free
Online
Unexplained Read/Write latency can be attributed to- cache hit-ratio, burst length, commands-in-a-row, AXI Bus arbitration, and video pipeline. During this Webinar, we will present the system-level modeling of complex video pipelines and their interface to memory using a Network of Buses. With this model, we shall optimize the address allocation to IP blocks, burst lengths and memory controller settings to get close to 98% efficiency.

We at Mirabilis Design have worked with a variety of architecture- processors, SoC, switches, multi-core and SSD controllers. We have cataloged our knowledge from these experiences to come up with a list of analysis, workload approaches and levels of abstractions. All of these are in a architecture exploration methodology that enables designers to gain extensive insight into the system operation.

Date : Wednesday, Feb 17, 2016
Time : 11 AM PST/ 2 PM EST
Duration : 1 Hour
Speakers : Deepak Shankar, CEO, Mirabilis Design Inc.
Case Studies : Darryl Koivisto, CTO, Mirabilis Design Inc.

 


comments powered by Disqus
Create your own event
Turn your passion into a business.
Join our mailing list