GW band gap in 2D system

[narzate]

Hi abinit users

I am getting different values for the GW band gap by running
version 7.10.5 and 8.4.2 of abinit for a 2D system with
cutoff in the Coulomb interaction.

I have checked the band gap results for bulk silicon and I am getting
the same results in both versions.

I have also run the test case of the isolated H chain calculation with
cutoff in the Coulomb interaction. This case is in the test directory
/abinit_dir/tests/v67mbpt/Input/t03.in
Here I also get the same GW band gap values in both versions.

Running the 2D case of MoS2 monolayer calculation "without
cutoff" in the Coulomb interaction in both versions of abinit,
I obtain the same values for the GW band gap.

However, when I run the 2D case of MoS2 monolayer calculation "with
cutoff" in the Coulomb interaction in both versions of abinit,
I obtain different values for the GW band gap.

Running the test file that is in attach I get the values:

Running version 7.10
E^0_gap 3.054
E^GW_gap 3.164
DeltaE^GW_gap 0.110

Running version 8.4
E^0_gap 3.054
E^GW_gap 5.029
DeltaE^GW_gap 1.975

What is it happening?
Am I missing something?

Thanks
Norberto