I am working on the ADSP21489 EZ-Kit.
The datasheet says that this processor is capable of working not only with 32-bit floating point formats but also with 40-bit as well.
I found the following discussion on a forum, but this discussion was quite some time ago:
So ;y auestion is -- in the case of our particular SHARC part (fourth generation) and the latest development tools, is it still true that we can only access 40-bit through assembly language, or can we access this with compiled languages now?
Thank you for your help,
Although support for 40-bit Arithmetic is not supported by the compiler, it is used by some run-time library functions and compiler support functions . 40-bit arithmetic can be enabled in C/C++ code by clearing the RND32bit on the MODE1 register but there are a number of factors that may result in arithmetic operations producing inconsistent results.
Refer to “Support for 40-bit Arithmetic” given in Page.No 305 of the VDSP++ Compiler Manual for more details . The link is given below.
It also describes the implications of using 40-bit arithmetic in C/C++ code.