Matt Venn
02/28/2024, 2:51 PMTim Edwards
02/28/2024, 3:25 PMTim Edwards
02/28/2024, 3:25 PMext2spice short resistor
)Mitch Bailey
02/28/2024, 3:26 PMMatt Venn
02/28/2024, 3:26 PMTim Edwards
02/28/2024, 3:27 PMconb
cell and tie the signal to LO
. (2) Add a metal resistor in the layout. Both of those solutions require matching components in both the netlist and layout. In particular, method (2) doesn't have any automation support and is a manual process.Roel Jordans
02/28/2024, 3:41 PMTim Edwards
02/28/2024, 3:47 PMconb
cell does in sky130, and that's what I recommend for digital signals. There is also a conb
cell in the HVL library that can be used for 3.3V high and low signals.Roel Jordans
02/28/2024, 4:12 PMTim Edwards
02/28/2024, 4:36 PMLarry Harris
02/28/2024, 6:51 PMMitch Bailey
02/28/2024, 7:02 PMio_out
signals which are actually inputs to the gpio cells. Leaving these unconnected can result in Hi-Z gate leaks in the caravel_core
region.
Interestingly, it appears that yosys will automatically connect unused output signals to ground when synthesizing (not with elaborating only). So if you have 2 synthesized macros with shared outputs, and one of the macros does not drive an output, your final signal will be shorted to ground. Your rtl simulation will work as expected, but not the gl simulation.carsten
02/28/2024, 8:39 PMMatt Venn
02/29/2024, 9:42 AMMatt Venn
02/29/2024, 9:42 AMMatt Venn
02/29/2024, 9:43 AMMatt Venn
02/29/2024, 9:44 AMMatt Venn
02/29/2024, 9:44 AMMatt Venn
02/29/2024, 9:45 AMMatt Venn
02/29/2024, 9:48 AMMitch Bailey
02/29/2024, 11:04 AMis a black box in the
in the lvs.report and Matching pins
and contains no devices
in the lvs.log. This will catch all the ‘X’ devices too, so you need to remove those from the output.Mitch Bailey
02/29/2024, 11:06 AMuio_oe[2]
pin in the source but not the layout? If so, you might have problems with LVS at the parent level.Matt Venn
02/29/2024, 1:17 PMMatt Venn
02/29/2024, 1:18 PM