Memory efficient finite volume schemes with twisted boundary conditions

In this paper we explore a finite volume renormalization scheme that combines three main ingredients: coupling based on the gradient flow, use of twisted boundary conditions and particular asymmetric geometry, for $SU(N)$ gauge theories consists hypercubic box size $l^2 \times (Nl)^2$, choice motivated by study independence in large $N$ theories. We argue has several advantages make it particularly suited precision determinations strong coupling, among them translational invariance, an analytic expansion reduced memory footprint with respect to standard simulations symmetric lattices, allowing more efficient current GPU clusters. test numerically determination $\Lambda$ parameter $SU(3)$ pure theory. show geometry no significant impact scaling violations, obtaining value $\Lambda_{\overline{MS}} \sqrt{8 t_0} =0.603(17)$ good agreement existing literature. The role topology freezing, is relevant applications, discussed detail.