DSP library for the ST10f269

Posted on April 07, 2005 at 10:08

Hello,

I don't know how to use the arithmetic functions of the DSP library for floating point operations. Is there a other way to realize floating point operations by the DSP library?

A software problem occurs when a floating point value in the C code is multiplied by 0 (without the DSP livbrary): then there is an exception, which is triggered by the ST10. Why is this happening?

Thank you for your support.

Posted on April 15, 2005 at 08:04

Hello,

The DSP library is only handling fixed point variables. For floating point, for many functions there are either algorithms available, either tools to generate ANSI C code that can be compiled with a ST10 toolchain.

For the handling of the exceptions in floating point libraries, this is related to the compiler. There is an option to check or not for these exceptions.

Posted on April 15, 2005 at 08:56

Hello,

I have to optimize a self-written C function, which uses floating point variables. Is this possible with the MAC unit of the ST10f269? How can I involve the MAC unit? Or are there other possibilities?

Where can I find these tools that generate ANSI C code for floating point operations? Is this code then already optimized?

Thank you for the information. :)

Posted on April 18, 2005 at 05:27

Quote:

How can I involve the MAC unit?

With the Cosmic ST10 compiler, there is at least one way to ''force'' the use of the MAC with C Code. However I don't think that floating point stuff is the best candidate to achieve significant optimization when using the MAC.

If you want mor details you can contact me offline.

Regards,

Luca (Cosmic)