Opamp test case using poor man's cascode

[soneryaldiz]

@srini229 , this PR adds one of the test circuits we discussed earlier. The test is currently failing:
2022-05-23 20:32:46 PnR.placer.Placer.setPlacementInfoFromJson ERROR : instance_name: X_I0 concrete_template_name: NMOS_4T_38150412_X2_Y1 not found.

[srini229]

@soneryaldiz do you want to use this PR to fix the failing test before merging?

[soneryaldiz]

Yes, let's fix if there is an issue. A variant, despite being generated, is not found by the placer. @854768750 , could you please take a look? To reproduce: pytest -vv -s -k opamp_poor

[854768750]

okay let me see

…

------------------ Original ------------------ From: ***@***.***>; Date: 2022年5月25日(星期三) 凌晨2:27 To: ***@***.***>; Cc: ***@***.***>; ***@***.***>; Subject: Re: [ALIGN-analoglayout/ALIGN-public] Opamp test case using poor man's cascode (PR #1115) Yes, let's fix if there is an issue. A variant, despite being generated, is not found by the placer. @854768750 , could you please take a look? To reproduce: pytest -vv -s -k opamp_poor — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[stevenmburns]

@soneryaldiz @854768750 This looks like the same problem as #1083 which is still pending. We might need a new start on it because what we have there now is worse than master.

[stevenmburns]

The issue is when the placement is read back into the router. To see this, try:

schematic2layout.py $ALIGN_HOME/tests/pdk/finfet_pdk/ckt_opamp_poor -p $ALIGN_HOME/align/pdk/finfet -l INFO -n 8 --flow_stop 3_pnr:gui
schematic2layout.py $ALIGN_HOME/tests/pdk/finfet_pdk/ckt_opamp_poor -p $ALIGN_HOME/align/pdk/finfet -l INFO -n 8 --flow_start 3_pnr:route

[854768750]

Yes, let's fix if there is an issue. A variant, despite being generated, is not found by the placer. @854768750 , could you please take a look? To reproduce: pytest -vv -s -k opamp_poor

@soneryaldiz This is the same issue as #1083

[srini229]

I tried to run this test after merging the changes from branch fix/multi_variant_placement_injection from #1083 to this branch. The test passes.

[soneryaldiz]

@stevenmburns and I decided to reduce variants to 1 to merge this PR. The known issue to be resolved in a future PR.

[srini229]

Looks good