Hello,
This requirement is a part of my thesis and I was wondering have anyone occurred to do a similar task of this sort.
I plan to implement NIOS II custom instructions that will accelerate the speed of an arbitrary precision integer arithmetic
(since parallel operations can be done in FPGA).
As a comparative study I am asked to compare the speed of a normal arbitrary precision integer library function with the custom instruction.
Say an ADDITION of 1028 bit + 1028 bit data. (This can be implemented in NIOS II custom instruction using verilog modules and can be called
from the application C/C++ program using macros in system.h file). The same addition can be done in the application code using a
multiple precision library as well.
Issue is with respect to porting such a library to the NIOS II build environment.
MPIR is such an arbitrary precision library which can be built with a Microsoft Visual C++ 2008 edition solution.
Is there someone who is experienced in bringing such libraries to the NIOS II IDE build environment?
If so that would be a big help.
Any pointers towards this direction will be appreciated.
Thank You,
Akhil Kalathungal
