Managing simulation time for divider circuit about fpu HOT 2 OPEN

Thanks for the help.the first methd of increasing the clock speed worked for us.It got reduced to 250 ns.

from fpu.

Hi,

The best solution will probably depend on your specific needs. I have a couple of suggestions that may help you.

1. Firstly, I'm not sure what clock speed you are running at, you might be able to increase the clock speed. Of course the maximum speed that you can clock the core will depend on the device you are targeting.

2. Secondly, this particular library was written with area in mind to keep the logic utilisation low. The amount of time that it takes to calculate a division is also variable. If you are trying to achieve a strict real-time performance requirement this may not be the best solution.

I have another project that might be a better fit. This divider in my verilog-math library is fully pipelined. It will use much more area, but will give you a much faster performance. The core has a fixed latency of 36 clock cycles, so if you ran at 100MHz (for instance) it would have a latency of 360ns from input to output, but would have a throughput rate of 1 division every 10ns. I'm not sure whether your 500ns requirement applies to the throughput rate or the latency, but this should give you sufficient performance in either case.

Thanks
Jon

from fpu.