Why isn't layernum=70, datatype=13 (aka met3:res) ...
# efabless
t
Why isn't layernum=70, datatype=13 (aka met3:res) listed in "gds_layers.csv"? It's used in "user_analog_project_wrapper.gds". Is it a mistake that it's not in "gds_layers.csv"? Or is a mistake that it's used in "user_analog_project_wrapper.gds"?
t
Layer:purpose 70:13 us a metal3 resistor. There is, unfortunately, rather a lot of confusion around layer:purpose values. The earlier DRC decks for the s8 process had purpose 15 = "short" and purpose 14 = "cut". I have never figured out what the difference is (or if there is any). There is a transistional document between s8 and sky130 that lists the generic purpose 13 = "type resistor (for diff, poly, wells, and metals)", and which does not list purposes 14 and 15 at all. Based on the transitional document, I implemented purpose 13 for all metal layers as "metal resistor".
As an identification layer that does not alter an actual mask layer, I don't worry about it. The "mistake" is that there isn't any consistency to any of the documentation I've been working with for this process. . .
t
Ok, so it should have been in the gds_layers.csv document. I figured.
What's the proper way of getting these things fixed? Make an issue on for the skywater-pdk repo on github?
t
Yeah, take a ticket and wait in line. I am considering making my own fork of skywater-pdk just so I can apply about a hundred patches that need doing. . .
s
In my past experience, the "cut" and "short" purposes have been useful occasionally where we needed to implement edits on hard macros without touching the block.
t
I didn't say they weren't useful, it's just that in SkyWater's interim document, they disappeared from the list of purpose types and were (apparently) replaced by the resistor purpose. Anything in that document could conceivably be in error.
s
ah