I'm not too sure about this one. Hopefully someone else can help you with this question.
I had the same thought as Karen... Not sure I understand the question. Could you rephrase the problem in another way?
Multiple gds2 files can be read by calibre DRC tool by SVRF commands LAYOUT PLACE CELL and Layout PRIMARY command.
Calibre tool combines the multiple gds2 into one top level gds2 and do the DRC verification.
Is there a way to output the combined gds2 created by the calibre tool.
We had about 50 gds2. We are able to run the DRC in one go using Layout place cell and layout primary commands mentioned in my first comment. But we see few errors in one of those cells. In order to view the error, we need the combined gds2 produced by calibre tool. I don't know how to get the combined gds2 created by the tool.
LAYOUT PLACE CELL is a new SVRF command introduced only from 2008_3 version of calibre.
Please let me know whether i conveyed the message clearly?
I see exactly what you mean now.
Calibre DRC is primarily a "checking" tool rather than "database manipulation". Merging 50 gds files together and outputting all the layers may be problematic in DRC.
Calibre DESIGNrev, especially with the help of a TCL script, is likely to be a much easier way to get the result you want. I don't have much direct experience with that but we may find related information or examples here in the Calibre Communities or maybe Supportnet.
If you only need a few layers for reference, and you don't mind that the hierarchy of the output may be different than the hierarchy of the input, you could use DRC CHECK MAP statements in the DRC run. These statements could send some reference data to a separate GDS output file that you could then use as overlay for your DRC error flags. This idea implies that you would also add a few rulechecks that used the COPY operation so that the drawn layers you were interested in would be output from the DRC.
Interested in your thoughts.
Thanks for your response.
If he only way we can view the constructed gds2 is through DRC CHECK MAP statements, then this particular command may not be of much help.
We can preserve the cell hierarchy by mentioning HCELL is the runset.
At least the tool should have the provision to list the cell placements. (Like cellA extents from (0,0) to (100,100), cell B extents from (125,125) to (225,225) etc.
Instead it just gives the total layout database extents.
I read about 100 cells in one go and listed all 100 cells as HCELL in my DRC run.
Three of my cells showed DRC errors. When i run individually the cells are DRC clean. So the error must be with my placement. But unfortunaley, i don't have any way to check my cell placements.
Do you have any feeback?
Maybe I could offer a small test case showing how DRC CHECK MAP statements can be used to create a separate output GDS file.
It would be a technique best suited to a situation where you knew of a few specific drawn layers associated with the few DRC errors you were seeing. For instance if the DRC errors related to metal spacing, then we could add a DRC CHECK MAP statement to copy the metal layer hierarchically to a GDS file. Then the metal spacing errors from the DRC output could be overlayed visually onto the special separate GDS file that contained the same hierarchical cell placements as constructed by Calibre for the DRC checks but the cells would contain just the metal.
If there were other DRC errors relating to diffusion spacing to poly, then we could add a DRC CHECK MAP statement for diffusion, and another for poly. Then that special GDS output file would contain altogether in hierarchical form, the cell placement information and layer information for metal, diffusion and poly.
Does that sound like something that would be helpful? Let me know if you would like a working testcase for this.
For a completely different approach, I suppose a script that could read the LAYOUT PLACE CELL statements from an SVRF rulefile and translate them to DESIGNrev statements to create a layout database would be great. I suppose PERL or TCL or AWK could be used to create something like that.