WORKAROUND (1):
Could you test these environment settings with Mir 0.26? The last one should in theory give you the same performance as you got in Mir 0.25 ...
MIR_CLIENT_FORCE_SWAP_INTERVAL=2 MIR_CLIENT_FORCE_SWAP_INTERVAL=3 MIR_CLIENT_FORCE_SWAP_INTERVAL=4
WORKAROUND (2):
Use only fullscreen RGBX clients and a server that supports bypass properly (mir_proving_server, not mir_demo_server). That will ensure the server process is never contending for GPU time and won't freeze or stutter.
WORKAROUND (1):
Could you test these environment settings with Mir 0.26? The last one should in theory give you the same performance as you got in Mir 0.25 ...
MIR_CLIENT_ FORCE_SWAP_ INTERVAL= 2 FORCE_SWAP_ INTERVAL= 3 FORCE_SWAP_ INTERVAL= 4
MIR_CLIENT_
MIR_CLIENT_
WORKAROUND (2):
Use only fullscreen RGBX clients and a server that supports bypass properly (mir_proving_ server, not mir_demo_server). That will ensure the server process is never contending for GPU time and won't freeze or stutter.