1 reply to this topic

[MEMS]

Hi I have written a C code which has float type variables.
when compile it for hardware if writes(i chose Altera platform) :

No floating-point library (double-precision) selected
.iMake: *** [Ashari.xhw] Error 1

the only way to solve this to chose Xilinx general VHDL
i can not use ultera platfomes. why?
what should i do if i want to use Altera???

2- is there any math library available for Impulse. i mean a library for matrix copmutions and other compilcated math functions.
as i am Electrical-Control eng i really need to work with matrixes but it will be very difficult and time consuming if want to write the basic functions for impulse enviroment.
thanx
MEMS

[RalphBodenner]

Floating-point is not yet supported for Altera platforms because free (as in $) IP cores are not yet available from Altera, as they are from Xilinx (included with price of CORE Generator). It is possible to use arbitrary floating-point cores with Impulse C by writing small pieces of HDL and XML that form an Impulse C library interface. See the application note IATAPP108 for details (download here: http://impulsec.com/...ation_notes.htm).

Several examples for complex math functions ship with CoDeveloper, such as CordicMath and SquareRoot. We also have a matrix multiply example that we hope to release with Version 3. The complex math functions in these examples can be called like C functions, or used in Impulse C processes.

We would like to add more such functions, and package them more cleanly in a library. Which matrix and arithmetic functions would be most useful to you?

Regards,
Ralph