KMLM List

[Sandeep S Srinivasa <sss@xxxxxxxxxxxxxxxxx>]

hi,
   Why dont you declare an event say "event1" and do 
"event1.notify(SC_ZERO_TIME)" in update_foo,
as well as a "wait(event1)" in issue_foo.
This will cause update_foo and issue_foo to be separated by one delta cycle.
regards
Sandeep


Sadagopan Srinivasan wrote:

> Hi,
>
>   I have two functions which are sensitive to positive edge of clock.
>
> void emif::update_foo()
> {
>
>   if(foo_var > 0)
>
>     foo_var--;
>
> }
>
>
>
> void emif::issue_foo()
> {
>
>  if(foo_var  == 0)
>
>    issue_command();
>
>
>
> }
>
> The implementation requirement is that update_foo should update in
> cycle X and issue_foo should see the change in cycle X+1. Since
> systemC has two phases of evaluation and update is it taken care of by
> itself normally or should I issue a dummy delay as follows.
>
> void emif::update_foo()
> {
>
>  wait(SC_ZERO_TIME);
>
>   if(foo_var > 0)
>
>     foo_var--;
>
> }
>
> this will delay the update of the foo till the end of cycle and so the
> issue_foo won't see the value until next cycle.
>
> Cheers,
>
> Sada
>
> _______________________________________________
> Systemc-forum mailing list
> Systemc-forum@xxxxxxxxxxx
> https://www.systemc.org/mailman/listinfo/systemc-forum
>
>
>  


--
			
			
		The glass is neither half-full, nor is it half-empty. 
	Nay, the glass itself is just twice as large as is necessary