I am currently using Cadence version IC126.96.36.1990.8. I am a regular user of Virtuoso Schematic Editor. However I don't have any experience related to automation in Schematic editor.
I am currently designing an IC that has around 15 - 20 IO Pads. Between each of these Pad's and internal logic (input or output) port, series of buffers (50-200) needs to be placed. I have the information related to the number of buffers corresponding to each Pad in an excels sheet. Doing this work manually is very boring and also tiring?
Would you kindly advice if its possible to automate the placement of buffers in the schematic between two signals based on the information provided via text file or .xls file.
Thanks and Best Regards,
Yes, this can be done. You need to first write some Skill code to read the text file using code like the following:
;Assume the txt file is a csv (comma separated values) dumped from an Excel spreadsheet looking like this:;portName,numBuffers
txtFileName="/tmp/my.csv"/*The text file for this example looks like this:port1,10port2,5port3,6*/procedure(myReadCsv(txtFileName) let((portTable filePtr lineData line) portTable = makeTable("portTable" nil) filePtr = infile(txtFileName) while(gets(line filePtr) lineData = parseString(line ",") portTable[car(lineData)] = evalstring(cadr(lineData)) ) close(filePtr) portTable ) ;let) ;proc
procedure(myBuildBufferSchematic(txtFileName) let((portTable cv x y dx inPinId outPinId buffId wireId bufInst bufWidth ) cv = geGetEditCellView() portTable = myReadCsv(txtFileName) x = y = 0 dx = 5.0 inPinId = dbOpenCellViewByType("basic" "ipin" "symbol") outPinId = dbOpenCellViewByType("basic" "opin" "symbol") buffId = dbOpenCellViewByType("myLib" "myBuf" "symbol") foreach(port portTable schCreatePin(cv inPinId port "input" nil x:y "R0") for(i 1 portTable[port] wireId=schCreateWire(cv "draw" "full" list(x:y x+dx:y) 0.05 0.05 0.0) bufInst = dbCreateInst(cv buffId nil x+dx:y "R0") bufWidth = abs(rightEdge(bufInst~>bBox) - leftEdge(bufInst~>bBox)) ;Subtract 0.05 so the wire overlaps the instance pin (adjust as needed for your buffer) x = x + dx + bufWidth - 0.05 ) wireId=schCreateWire(cv "draw" "full" list(x:y x+dx:y) 0.05 0.05 0.0) schCreatePin(cv outPinId strcat(port "_out") "output" nil x+dx:y "R0") y=y+1 x=0 ) printf("Finished\n") ) ;let) ;proc
;lbl = schCreateWireLabel(cv car(wireId) x:y+0.05 labelName "lowerCenter" "R0" "stick" 0.05 nil)
Your needs will certainly vary from what I've provided, but here are some things to note:
I didn't name the wires between the buffers, but if you need to you can use the schCreateWireLabel command shown above.
Generally this is possible.- Output a csv file form Excel.- Read it in with SKILL file IO commands.- Post processes in SKILL- Do a schematic placement with db* commands e.g. dbCreateInst
Thanks Derek for your detailed reply and also the sample script. That's very helpful.
I followed all the steps that you have mentioned above.
1. So first I created a sample csv file (myfile.csv) with following data:
In CIW I entered
The copied your SKILL script to a .il file (build_buffer.il) and added the buffer library and cell name.
Then in CIW I did
However I am getting an error:
*Error* evalstring: argument #1 should be a string (type template = "ts") - nil
Is it something to do with "portTable[car(lineData)] = evalstring(cadr(lineData)) ", since the buffer numbers in the txt file are not strings??
I'm guessing there is probably a blank line in the text file. The evalstring is going to convert the string to a number when it stores it in the table. Make sure there are no trailing blank lines in the file.
You could error check for that this way: while(gets(line filePtr) lineData = parseString(line ",") when(length(lineData)==2 portTable[car(lineData)] = evalstring(cadr(lineData)) ) )
Also, if you are going to write Skill, the following commands will give you even more detail when your Skill fails:sstatus( stacktraceDump t )sstatus( stacktrace 8 )sstatus(traceArgs t)
Thanks again Derek. You were right, there were trailing blank lines in my text file.
Now the script runs perfectly.
Also thanks for the tips related to debugging.
That was indeed very helpful !