Building string variables from string variables in ADE

If you're using ADE Explorer/Assembler, then you can have the strcat in the design variable or global variable field. I think that does what you want. You could also put $PROJECT/load_waveforms in the include path, and then you only need to build the string with the actual file name part.

Andrew

The include path is a good point. Thanks, Andrew!

I am a bit puzzled with strcat in ADE, because in some cases it has worked, but in others it did not, and I am not sure what made the difference...
Below is a simplified testcase for strcat() where all but the variable "path" is resolved (unresolved variables are bolded). Do you see/know what is the diffrence, i.e. what is needed to resolve a variable win strcat()'s arguments?

parameters my_strcat="foobar" my_strcat3="/a/b/c/efd/text.txt" \
 my_cellview_name="foo_bar_barrr" pwl_cellview="onceUponATime" \
 my_cellviewname="foooobar" \
 path="$PROJECT/oa/foo/bar/foobar/onceUponATime/text.txt" \
 my_strcat2=strcat(my_strcat "x/x" "apple" "x/x" "pear" ".csv") \
 my_strcat4=strcat("$PROJECT/oa/foo/bar/foobar/" my_cellview_name \
 "/text.txt") path2=strcat("$PROJECT/oa/foo/bar/foobar/" \
 my_cellview_name "/text.txt")

The corresponding ADE setup: 

PS: note for the future: strcat is resolved after netlisting, i.e. Recreating the netlist without simulation would put the whole strcat() expression, and not just the result into the parameter in the input.scs file.

Dear Zol,

Zol said:

I am a bit puzzled with strcat in ADE, because in some cases it has worked, but in others it did not,

From your description, it is working in that the simulation runs successfully, but you notice that some string parameters shown in the input.scs file are shown as text strings (i.e., without any references to other strings) and other string parameters include a strcat() function with one or more of its arguments a name of a string parameter.

Zol said:

Below is a simplified testcase for strcat() where all but the variable "path" is resolved (unresolved variables are bolded). Do you see/know what is the diffrence, i.e. what is needed to resolve a variable win strcat()'s arguments?

Are your variables  mix of global and local design variables? There may be a difference in how they are passed to the input.scs.

Shawn

Thanks for the info - I just recreated your design variables (thanks, you were clear that these were design variables, not global variables, although that should work too. Not sure why Shawn asked that as it was clear). I tried using IC6.1.8 ISR33 (IC6.1.8-64b.500.33) and in the simulation netlist, all the strcat were resolved:

If you do Simulation→Netlist→Recreate then the strcat are not evaluated - the reason being is that this produces a single netlist and your expressions may be dependent upon variables that are swept - so it doesn't attempt to evaluate them when creating the netlist that way - not an issue since that's never used for simulation). Maybe that's what you meant by your PS in the above post.

Since your netlist has a mix of expanded strcat and unexpanded, my guess is that you're using an older version. There have been various fixes to expression evaluation within the lifetime of IC6.1.8/ICADVM20.1 so I would recommend you use the latest hotfix if you can.

Note that simulation will succeed if there are some strcat() remaining in the netlist, provided that the parameter which suffers from this is not used. Spectre itself won't evaluate the strcat itself - but any unused parameter (or chain of unused parameters) is ignored.

Andrew

Thanks for the info - I just recreated your design variables (thanks, you were clear that these were design variables, not global variables, although that should work too. Not sure why Shawn asked that as it was clear). I tried using IC6.1.8 ISR33 (IC6.1.8-64b.500.33) and in the simulation netlist, all the strcat were resolved:

If you do Simulation→Netlist→Recreate then the strcat are not evaluated - the reason being is that this produces a single netlist and your expressions may be dependent upon variables that are swept - so it doesn't attempt to evaluate them when creating the netlist that way - not an issue since that's never used for simulation). Maybe that's what you meant by your PS in the above post.

Since your netlist has a mix of expanded strcat and unexpanded, my guess is that you're using an older version. There have been various fixes to expression evaluation within the lifetime of IC6.1.8/ICADVM20.1 so I would recommend you use the latest hotfix if you can.

Note that simulation will succeed if there are some strcat() remaining in the netlist, provided that the parameter which suffers from this is not used. Spectre itself won't evaluate the strcat itself - but any unused parameter (or chain of unused parameters) is ignored.

Andrew