Closed tianyma closed 3 years ago
Cycle counts definitely longer for mul (pcpi_valid for ~37 clocks) vs fast_mul (~3 clocks) here - on verilator and on the fpga target - perhaps post a snippet of your verilog and I might be able to help.
Cycle counts definitely longer for mul (pcpi_valid for ~37 clocks) vs fast_mul (~3 clocks) here - on verilator and on the fpga target - perhaps post a snippet of your verilog and I might be able to help.
Thank you, it's my mistake, I set the wrong parameter, and the two tests actually both use fast mul. Now it works.
Hi, maybe this is a basic question, I am new in IC design. I found that in function simulation, norm_mul and fast_mul have the same cycle counts, which both are 6 cycles, and I cannot figure it out. Can anyone explain it for me?
I use iverilog to compile and gtkwave to view the wave, and I see the signal
are both 6 cycles for either normal multiple or fast multiple.