[Pw_forum] B3LYP - (k+q) points

[Alexandra Davila]

Dear QE users,

I am trying to quantify the difference between PBE and B3LYP for Cl/Au 
adsorption energy.
My system has 13 atoms, c(2x2) surface unit cell with 6 layers and 12 
x12 x1 kpoints  ( 43 irreducible kpoints, I know, that it is too much).
The pseudopotential is ultrasoft generated with BLYP functional.

In my input file, I set the following flags:

  input_dft='B3LYP'
  nqx1=2, nqx2=2, nqx3=1

I performed first a scf calculation without the flags above, then one 
with (which takes a lot of time and even now it is not ready).

My question is about the (k+q) points, in my output it is written:
EXX: setup a grid of 128 q-points centered on each k-point
      (k+q)-points:

I would like to understand from where does the number come? As far as I 
understood, (k+q ) points are in the set of k points, why I have more?
What can I do to reduce the cpu time?

Thanks for your time and patience,