cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
chenhai
Visitor
Visitor
7,373 Views
Registered: ‎02-14-2008

Problem regarding to accumulation and hardware co-simulation

Hi all,
 
I need an accumulation in my design. However, it doesn't work in hardware co-simulation.
 
This is my situation:
1. [1:10;[11:20]]' from workspace as the input. Process by an default setting accumulation. Finally output to a scope. It works very well in the matlab simulation case.
 
2. Generate the accumulation to LIB with the point2point network cosimulation. The result return is unexpectable, very different from that of the software simulation.
 
Would anyone tell what happens to this, pls? How to get the correct answer in hardware co-simulation? Thanks.
0 Kudos
3 Replies
chenhai
Visitor
Visitor
7,363 Views
Registered: ‎02-14-2008

It seems that 8-bit input and 8-bit output can get the correct result.
But 8-bit input and 16-bit output can't.
I will try again today and hopefully can make it clear.
0 Kudos
chenhai
Visitor
Visitor
7,337 Views
Registered: ‎02-14-2008

It works after setting the "clock source" as "single stepped".
 
0 Kudos
jeffreyh
Xilinx Employee
Xilinx Employee
7,268 Views
Registered: ‎08-07-2007

If you're not using "single stepped" mode the hardware will not be synchronized to the software and the data captured and passed to the hardware will be somwhat random since the hardware co-sim interface is not able to run at the full board system rate typically.
0 Kudos