create a pCell with m-factor

Hi David,

If you're using VLS XL, you don't really need to do anything. You can have the parameter m in the CDF (and passed to the simulator as an m-factor) and VLS XL will automatically generate that number of instances in parallel. You can then place them, abut them, use them in modgens as you see fit. There are controls in VLS XL to define the parameter used to represent m, but the default includes the parameter name m.

Coding inside the PCell itself is probably not the best option as then it removes the flexibility to place or interleave devices to improve matching.

Regards,

Andrew.

Hi David,

If you're using VLS XL, you don't really need to do anything. You can have the parameter m in the CDF (and passed to the simulator as an m-factor) and VLS XL will automatically generate that number of instances in parallel. You can then place them, abut them, use them in modgens as you see fit. There are controls in VLS XL to define the parameter used to represent m, but the default includes the parameter name m.

Coding inside the PCell itself is probably not the best option as then it removes the flexibility to place or interleave devices to improve matching.

Regards,

Andrew.

wow, so straightforward. almost too good to believe. thanks a lot. 

hi Andrew,

a following up question: I had this cell with "m" parameter and behavioral schematic model, Spectre complains that "m" should be reserved and so ignored. Is there a way not to ignore that? seems my behavioral VerilogA model still works with "m" parameter. 

     Warning from spectre in `cell', during circuit read-in.
     WARNING (SFE-57): "input.scs" 55: Parameter name `m' is reserved for the
     subcircuit multiplicity factor. So a formal parameter of that name is ignored. 

thanks,

David

Hi David,

With spectre, setting m on a hierarchical instance automatically passes that m to all the components underneath - multiplying it with any m on those components. The result is that m ends up on the leaf devices and is handled there (you have to specifically give a special attribute on the definition of m in veriloga to have that pick up the m factor).

i’m guessing that you’re using pPar to pass the m-factor through the hierarchy - you don’t need to do that, and it’s probably that which is causing the message. 

This is a bit of guesswork as I can’t see your setup. Perhaps it would be best to go via customer support so we can look at your data?

Regards,

Andrew

hi Andrew,

in behavioral schematic, yes I had pPar("m") and it's working now after I remove pPar("m").

in veriloga model, I have parameter m and apply it to the code as needed. 

working great now. 

thanks,

David

hi Andrew,

one more question again although it's about Calibre: when I run LVS by Calibre, it doesn't recognize the m-factor we put in here so it sees less instances in schematic than in Layout thus fails LVS. 

is there something we can do here for Calibre to recognize the m-factor?

thanks,

David

First question - are you seeing m in the CDL netlist?

Andrew

hi Andrew, 

quick answer is no. I don't see m parameter from LVS/cell_extract.src.net. 

maybe worth mentioning this cell is metal based custom design so I have “LVS BOX unitcap” configured. would this mask the m-factor off?

thanks,

David

For the purpose of LVS netlisting (I.e. CDL netlisting) does this end up as a simple instance in the CDL netlist or does a subckt get output ? If there’s an auCdl view of the component, it’s probably just a matter of adding “m” into the instParameters section of the auCdl simulation information in the Base CDF for the component. If that doesn’t work please share what the auCdl sim info looks like in the CDF and whether you have an auCdl stopping view for the cell. 

I cant answer about “LVS BOX unitcap” because I don’t have access to Calibre (a Mentor product).

Regards,

Andrew

hi Andrew,

just feedback after we tried a couple things. after we created new marker layers to recognize the cell in LVS without blackbox, we can see the Multiplier now so no worry.

thanks,

David