Dear Hongyang,
As far as I know, a viable way in CP2K is really to do Gamma-point calculations in big supercells. This may look to be expensive but CP2K offers the TC-LRC and the ADMM techniques; see especially
[M. Guidon, J. Hutter, J. VandeVondele: J. Chem. Theory Comput. 5, 3010 (2009)],
[M. Guidon, J. Hutter, J. VandeVondele: J. Chem. Theory Comput. 6, 2348 (2010)].
I recommend you to have a look also at several exercises and examples on the CP2K site. In my jobs, hybrid-DFT calculations with these techniques were running about 2 times slower only comparing to PBE calculations with the same basis sets.
I know also about papers
[C. M. Yim et al.: Phys. Rev. Lett. 117, 116402 (2016) - see also its supplement. info],
[A. R. Elmaslmane, M. B. Watkins, K.P. McKenna: J. Chem. Theory Comput. 14, 3740 (2018)],
[J.C. Carey, K.P. McKenna: J. Phys. Chem. 122, 27540 (2018)]
which you may find very convenient.
Best regards
Martin.