If you experience any such performance problems then please open a bug by running:
ubuntu-bug gnome-shell
and keep the discussion in the bug.
P.S. I noticed the Ubuntu default extensions are causing some noticeable performance loss so you might want to try running without those. Either install gnome-session or you can disable the extensions by installing gnome-shell-extension-prefs.
Ok, I will report the bug. Sorry if this wasnāt the right place.
And, indeed, the gnome-session seems to be a bit smoother.
Thanks for the bug report. It seems to be the first and only performance bug report against Gnome Shell proper in 20.04 
Gnome Shell is in great shape in 20.04! I think I noticed this bug because everything else is so good 
Thank you Daniel and all the team for the superb job!
I never mentioned this before, but I routinely check @vanvugt GitLab profile because I know heās always working on something to make the desktop experience better. Great stuff the last two and a half years.
Keep it up!
Agreed! Iāve been wondering what happened to @jbicha ? They both were working on Ubuntu Desktop but havenāt heard much from Jeremy the last couple of cycles. Does he still work on Ubuntu?
Iāve reported severe 20.04 performance regressions above, in particular for the overview animation in X11 and when comparing a vanilla Wayland session vs Fedora 32 and Arch. All this in a Huawei Mate book Pro X (Intel 620 + NVidia). Now for the silver lining: using kernels above 5.5 (Iām on 5.6.4 now) put performance on par what Iāve seen in Arch and Fedora (both shipping 5.6), indeed it feels smoother in Ubuntu for X11. Another fine tuning I did was to turn off Intel Panel Self Refresh (psr), the improvement was not as spectacular as the one the kernel upgrade brought, but it was an improvement anyway. Sadly 5.6 is not the LTS kernel, but fortunately itās easy to keep a number of kernels installed.
@memeplex, thanks for the bug report. I am keen to see your problems solved but canāt make any guarantees about how long it will take. Regardless, please keep any comments relevant to that problem in the upstream bug to avoid duplication.
Sure @vanvugt, I know itās not easy to compile the shell and perform benchmarks myself, but I can dedicate a couple of days to this if you need me to collect detailed information. But letās keep it upstream as you said. Thanks!