search

evanberkowitz / supervillain

MCMC for modified Villain constructions
https://supervillain.readthedocs.io/
Other
0 stars 0 forks source link

Worldline Worm Spin_Spin #107

Closed evanberkowitz closed 1 month ago

evanberkowitz commented 8 months ago

If we can measure the Spin_Spin correlator during a worm update, we can include the SpinSusceptibility+Scaled observables in the autocorrelation time without encountering bad costs as described in #102.

evanberkowitz commented 1 month ago

As we learn in #136, if we make inline measurements using the worm, the susceptibilities actually need to be normalized in expectation value (after averaging) and therefore need to be DerivedQuantities, which cannot be used in autocorrelation time anyway.