Honored Contributor
Honored ContributorHello All
I make multicycle multiplication instruction and for implementation this instruction need ~ 110 LE's - it work up to 3 times faster then usual multiplication without hardware support and it not decrease Fmax - because it have verry little LE's amount A -32 bitwidth B -32 bitwidth result - 32 bitwidth almost all elements used for data latching But it work only in my program where i directly call this instruction http://forum.niosforum.com/work2/style_emoticons/<#EMO_DIR#>/sad.gif Is it possible to use my mul instruction in C C++ compliator like in NIOS5.1 when i select hardware multiply by Logic elements ? I menan if i write in C++ Y=X*Z; - my intruction calc this. If i do it my program will work murch more faster then before The same for division - i can make very little multicycle instuction but using in standart compilator will not be able. I think if name of my Mul instruction will be the same like in standart SOPC support may be possible to just enable somehow compilation with using hardware multiply It will speed up any nios2 system without big growing numberic of LE's In my opinion it must be very easy need to correct some parameters in NIOS2 generation scripts to eable using hardware support of multiplication in C++ compilation If some one know how to do it - help with it, please
