We are doing some bench marks using Cyclone and Cyclone II doing multiplies. We have written some code that will do 1 million multiplies and return the amount of time it took to perform the task. Our results for the Cyclone II are not as good as expected.(we are using the CycloneII eval board, and a custom Cyclone board) I have compiled the CycloneII core using NIOSII/f using it Embedded Multipiers.
Is there something that I must to in software to be sure that it is using the Embedded Multipliers?
Is there something in the core that I may have done incorrectly?
Maybe I am getting its best perfomance. With the CycloneII it is taking about 9 seconds to perform the task where the Cyclone is taking a little over 20 seconds. I would have thought the Cyclone II would have been better.
Any help would be GREATLY appreciated!
Brandon
