Fig. 10.

Effect of employing approximate SHTs in the preconditioner. To simulate approximate SHTs, for every SHT computer, we have multiplied each coefficient of the result with a random variate drawn uniformly from the interval [1 − ϵ, 1 + ϵ], simulating numerical noise at the ϵ-level. Note that we carried out these simulations on a resolution of Nside = 128, using the downgrade procedure outlined in Appendix C. The experiment is otherwise identical to the one in Fig. 9. The approximate SHTs are employed both in the pseudo-inverse preconditioner and the mask-restricted multi-grid solver, while the (non-preconditioner) matrix-vector multiplication uses the full accuracy provided by Libsharp.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.