In the

.mag

file, is

FIXED_BBOX

in the same units as

MASKHINTS_*

?!?

Yes, it should be. Both are in internal units (which can vary, so the units for both are dependent on the

magscale

value at the top of the file. But

FIXED_BBOX

and

MASKHINTS

units should always agree with each other).

That doesn't seem to be the case. Because they are

string FIXED_BBOX 0 0 210 70
string MASKHINTS_NSDM 84 0 394 140

But yet the generated GDS has the correct height for both (same height for prBoundary and nsdm)

Are there any other properties in the cell?

Just

FIXED_BBOX

and

MASKHINTS_NSDM

Congratulations, you have found a bug. : )

🤣

"_*Zarro Boogs Found"*_

Yes, I'm aware of where it occurs and what needs to be fixed. MASKHINTS_* must be treated like FIXED_BBOX.

Ah ok, the issue is the maxhints not being scaled, not fixed_bbox being scaled, I see 😅

It got the way it is because when using the default

sky130A.magicrc

, the database is scaled down by a factor of 2 to the manufacturing minimum grid (5nm), and most layout files end up at that scale. That is, I think most existing .mag files have

magscale 1 2

and the

MASKHINTS

have the same scale as

FIXED_BBOX

. I don't think it will be too disruptive to fix it.

Weird because AFAIK I'm using the default sky130A, I haven't messed with anything and yet I don't have

magscale

in my files and obviously any fix would break them.

I have fixed the code for magic (on opencircuitdesign.com; as usual, it mirrors to github in about 12 hours). I'm afraid there really isn't any good solution except to fix the problem. As previously written, it is impossible to determine what units the MASKHINTS are in, in the .mag file. The only units that are meaningful in the .mag file are lambda units (or a multiple thereof, given a non-default "magscale" line). Internal units can get changed arbitrarily.

Ack. I just updated magic, fixed up all my

.mag

files manually and regenerated the GDS and they match the previous result.