|
||||||||||
Sonics Improves NoC Concurrency Management for SoC Designs with Multi-Channel Memory Sub-systems, Addresses Place & Route Tool RestrictionsSonicsGN 3.0 Expands Patented IMT to Exploit Transaction Parallelism and Increases User Control of Hierarchical RTL Partitioning to Guide Physical Design Milpitas, Calif. – June 23, 2015 – Sonics, Inc., the world’s foremost supplier of on-chip network (NoC) technologies and services, today released SonicsGN® 3.0, the latest version of the company’s flagship NoC. SonicsGN 3.0 expands on the interleaved multi-channel technology (IMT) that has been patented and proven in SonicsSX® and includes new layout optimization features for design flows based on modern physical synthesis and place & route tools. SoC architects using SonicsGN 3.0 can eliminate multi-channel DRAM access bottlenecks with IMT and reduce layout tool iterations with flexible user control of hierarchical RTL partitioning and re-timing stage insertion. “We originally patented IMT in 2008 and it has been routinely implemented by many of our SonicsSX® customers. At that time, SoC designers were early in the move to multi-channel memory architectures,” said Drew Wingard, CTO of Sonics. “Use of multi-channel DRAMs and memory sub-systems is now pervasive in SoC designs, for example, in mobile applications where maximizing memory throughput takes precedence over increasing memory capacity. To address this trend, we’ve expanded IMT in SonicsGN 3.0 to give SoC architects much more effective management of transaction concurrency and memory sub-system performance in their designs. At the same time, we’ve enhanced SonicsGN’s handling of hierarchical partitioning and re-timing to provide better layout guidance when designers generate their RTL netlist for physical synthesis and detailed place & route.” Concurrency Management for More Efficient DRAM Access and Transaction Ordering Another related concurrency management problem is ordering of requests that may be simultaneously outstanding to multiple targets, including multi-channel memories. SonicsGN’s flexible reordering buffer architecture enables single initiator agents to have transactions outstanding to an arbitrary collection of targets, while respecting the protocol-defined ordering. SoC designers using the reordering buffer can further enhance concurrency by creating separate transaction tags where none existed or by restricting the reordering behavior. Layout-Optimized RTL Netlist Generation Availability About Sonics, Inc.
|
Home | Feedback | Register | Site Map |
All material on this site Copyright © 2017 Design And Reuse S.A. All rights reserved. |