Comments on: What embedded development tool do you want?

By: B.E. @ TI

B.E. @ TI — Fri, 30 Dec 2011 16:48:14 +0000

Great article and read Robert and have a Happy Holiday!

By: one and zero @ TI

one and zero @ TI — Fri, 30 Dec 2011 16:47:55 +0000

Hi Robert,

great article … here’s my answer to your question:

Code Composer Studio v5.1 is great for all customers using TI products from low power MSP430 up to high performance C6000 multicore architectures. CCS is based on the latest version of Eclipse with no modifications. This allows customers to integrate a wide range of 3rd party plugins and allows upgrades of open source components of CCS. See a introduction video here:

software-dl.ti.com/…/CCSv5GettingStarted.html

CToolsLib provides a set of fundamental libraries for the configuration and control of debug and profiling trace modules common to high performance C6000 multicore and OMAP devices. See more details here:

processors.wiki.ti.com/…/CToolsLib_Article

And there are advanced Emulators that allow you to perform Program Counter Trace which is natively supported in CCS and supplemented by the CToolsLib functionality:

processors.wiki.ti.com/…/Debugging_With_Trace

Kind regards,

one and zero

By: Paul A. Clayton

Paul A. Clayton — Tue, 20 Dec 2011 08:13:08 +0000

This feature may already be present in existing tools, but it seems that it would be useful to have a software compilation system that could prove the equivalence of a simple, easily understood encoding and a more obscure, highly optimized form. (This would seem to require support for preconditions and the like since many optimizations are dependent on special knowledge that is not expressed in standard C source code.)

Generating a specification based in the English language which can be used for formal verification is very likely impractical, but generating a correct simple C-like language specification and proving equivalence may be practical.

Along similar lines libraries and portable code might benefit from a means of providing multiple ‘overloadable’ implementations for which the compiler could prove equivalence and from which the compiler could choose an optimal implementation for a given target use and goals, ISA, processor microarchitecture, and system features. Ideally, the tools could even suggest updates of a variant implementation when a change in the canonical implementation breaks the variant (as well as detecting the breaking), possibly even feeding back to the developer that a specific small change will ‘break’ many variants so that the developer can make a more informed decision about whether the change is worthwhile or an alternative is appropriate. Such a system might make maintaining code with many optimizing variations practical.

(I am just a thinker not even a programmer so the above might not be practical or useful.)