editor's blog
Subscribe Now

Synthesizing TLM Models

Architectural exploration and design implementation all too often are two separate tasks implemented by two completely separate groups. Once a high-level TLM model has been tested and approved, it goes on the shelf while a designer starts from scratch to generate a synthesizable design.

While this may sound simply wasteful, things aren’t so simple. TLM models are abstract, using busses and transactions. An RTL design has to specific signals at the individual level – the TLM model doesn’t have that, so the TLM model tends not to be particularly useful as a starting see for implementation.

Calypto is trying to address that, as they described in a discussion at DVcon. They’re making available a set of interfaces that can be used at multiple levels of abstraction – and, specifically, which can be used to go from architecture to synthesis. They’re starting with AXI, but plan others.

The tie-in with Mentor here is obvious. First off, you may remember that they bought Mentor’s Catapult C. So they have a stake in the high-level synthesis game now. Second, Mentor has a technology that they call Multi-View that allows a single model to be used and viewed with different levels of abstraction for different purposes.

These technologies converge in this IP announcement. You can find more in their release

Leave a Reply

featured blogs
Sep 20, 2019
Earlier this year, on July 20, we celebrated the 50th anniversary of the first landing on the moon. I wrote about it'€”well, mostly about the Apollo Guidance Computer'€”that day in my post The First... [[ Click on the title to access the full blog on the Cadence Communit...
Sep 19, 2019
The word '€œrevolution'€ is very common in today'€™s world.  Whether it'€™s a revolutionary new smartphone or a revolution in personal transport, it is a phrase that is used frequently.  How many of these things are truly groundbreaking, or how many are vic...