editor's blog
Subscribe Now

Managing Multicore Tasks

We’ve looked in the past at some of the APIs put together by the Multicore Association, notably the MCAPI communications API (and its update) and the MRAPI resource management API. In the same spirit, they are now taking up the issue of task management.

The whole idea with multicore software is that a program can be decomposed into multiple pieces that can be run concurrently. How you do that partitioning can be a hard problem, and is receiving some tools attention as the multicore design process gets some automation. You might think of these pieces as threads, but, in fact, threads are typically a specific implementation in SMP systems. Alternative parallel configurations can include multiple independent programs (rather than threads), each with its own OS, or even simple run-to-completion programs on “bare metal,” i.e., with no OS (and, hence, no threading services) at all.

So rather than getting bogged down in worrying about whether these “pieces” of the program are threads or programs or whatever, they can be generically referred to as tasks. And, while SMP OSes can handle the management and scheduling of threads, there’s no general task management solution for non-SMP systems – or, more critically, there’s no general approach that works for both SMP and AMP, homogeneous and heterogeneous systems.

Even where threading services exist, they can have far too much overhead for many embedded programs. Including the cost of creating and destroying threads, the overhead can completely wipe out any theoretical gains that the parallel version was supposed to provide. If your program exploits fine-grained parallelism, with hundreds or more small tasks, then the thread management can take longer than the actual execution of the thread itself.

In order to address this, the Multicore Association is taking up the creation of a task management API, called MTAPI, in order to provide a general approach for all architectures, and one whose implementation can be tailored to limited resources as needed. The process is just starting, and they’re soliciting input and participants.

More info can be found in their release

Leave a Reply

featured blogs
Oct 16, 2019
In this week's Whiteboard Wednesdays video, David Peña discusses Cadence'€™s focus on models for various emerging memory standards. https://youtu.be/_Xps6I6kE0E [[ Click on the title to access the full blog on the Cadence Community site. ]]...
Oct 15, 2019
As technology advances, it's becoming harder and harder to know what is real and what isn't....
Oct 14, 2019
My working life includes a lot of writing – blogs, articles, conference papers and white papers are typical of what I produce. A common factor of my writing is that it is aimed to be technical and instructive. What I do not like writing is sales pitches. I can accept th...
Oct 14, 2019
In 1995, I attended a seminar in which the presenter told us that copper was dead.  This sort of statement is not new. The connector market is filled with armchair pundits who predict the demise of everything from D-Subminiature connectors (which are very much alive and ...
Oct 11, 2019
[From the last episode: We looked at subroutines in computer programs.] We saw a couple weeks ago that some memories are big, but slow (flash memory). Others are fast, but not so big '€“ and they'€™re power-hungry to boot (SRAM). This sets up an interesting problem. When ...