Open wright-stuff opened 9 months ago
Using Windows
# daemon env1 conda create -n bad_daemons yaqd-core python=3.12
# daemon env2 conda create -n good_daemons yaqd-core python=3.11
if I initiate a client call many times:
for _ in range(10**7): c.get_position()
a daemon in env2 keeps the memory size to 10s of MB, but an env1 daemon balloons to ~10 GB
The problem is quite practical; running yaqc-qtpy on one of the Wright-Group computers can make daemon memory reach 10s of GB after a few hours.
Probably an upstream issue that will be fixed, but be advised: avoid python 3.12 for daemon environments.
Using Windows
if I initiate a client call many times:
a daemon in env2 keeps the memory size to 10s of MB, but an env1 daemon balloons to ~10 GB
The problem is quite practical; running yaqc-qtpy on one of the Wright-Group computers can make daemon memory reach 10s of GB after a few hours.
Probably an upstream issue that will be fixed, but be advised: avoid python 3.12 for daemon environments.