How do you write a level-shifter strategy with set_level_shifter, and what do its key options control?

Short Answer

set_level_shifter (UPF) tells the tool which ports of a domain get level shifters, for which voltage direction, and in which domain the cell sits. Low-to-high crossings always need a shifter; high-to-low ones depend on the library and the voltage gap, which is what -rule and -threshold let you filter. Without any strategy, ICC2 still inserts shifters from the voltages in the power states.

Technical Explanation

• -domain names the domain whose ports are shifted; copy the option name exactly, since a misspelled variant fails to load.
• -applies_to inputs, outputs or both, narrowed with -elements or filtered with -source and -sink.
• -rule low_to_high, high_to_low or both (default both) keeps the strategy to one direction.
• -threshold applies the strategy only where the voltage difference exceeds the value; the 1801-2015 default is 0.
• -location: 1801-2015 defaults to self, while the ICC2 MV guide lists automatic as default. Set it explicitly.
• -input_supply and -output_supply name the supply sets for each side; -no_shift blocks insertion on listed ports.
• You cannot combine -location with -update in ICC2, so fix the location when you first write the strategy.

# [UPF]  mychip.upf
set_level_shifter FROM_PD_CPU_LST -domain PD_CPU -applies_to outputs -rule low_to_high -location parent
set_level_shifter TO_PD_CPU_LST -domain PD_CPU -applies_to inputs -rule high_to_low -location self
# [ICC2]  icc2_shell
create_mv_cells
check_mv_design
analyze_mv_design -level_shifter -through U_CPU/dout

What To Check

• Every PD_CPU output into a 1.0 V domain has a low-to-high shifter.
• High-to-low crossings are covered or deliberately excluded with a library-backed reason.
• Each strategy states -location explicitly.
• The chosen location has both supplies available.

Healthy, Suspicious & Hard-stop Results

• Healthy (illustrative): check_mv_design reports no missing level shifters, and every PD_CPU output shows an up-shifter.
• Suspicious (illustrative): A high-to-low crossing has no shifter and no written reason for skipping it.
• Hard stop: A 0.9 V to 1.0 V output drives PD_MYCHIP with no shifter, or load_upf rejects a misspelled -domain option.

What The Interviewer Is Testing

• Do you know low-to-high is mandatory and high-to-low is conditional?
• Can you explain -rule and -threshold as filters rather than cell choices?

Follow-up Question & Model Response

"Why is a low-to-high shifter mandatory while high-to-low sometimes is not?"

Candidate Model Response: A 0.9 V high driving a 1.0 V gate may not turn its PMOS fully off, so the receiver leaks and switches slowly. Going the other way, a 1.0 V signal into a 0.9 V gate turns it fully on and off. Whether the input can tolerate the higher voltage is a library and reliability question. So you skip high-to-low shifters only when the library says the cell is safe at that gap.