Free Access

Table 4

Gain in time achieved by “source loop” parallelization of the “data-space” algorithm.

Algorithm Average time (s) per “source loop”
Dataset version Number of threads
1 2 4 8 16 32

V0332+53 Full 22.8 15.7 10.1 8.1 7.7 5.2
(391 exposures) limited 3.1 2.0 1.8 1.1 1.0 0.7

GX 339-4 Full 607
(1183 exposures) limited 119 43 30 23

Notes.The “limited” version is parameterized as indicated in Sect. 6.3.4. The dataset characteristics are detailed in Table 3. A computer with 64 (2.62 GHz) processors was used. The total number of “source loops” performed for the “full” version was 519 and 2359 for V0332+53 and GX 339-4. Similarly, 508 and 2296 loops were performed with the “limited memory” version. The total time spent in the “source loop” for the “full” (“limited memory”) “data-space” algorithm was obtained by multiplying the average single-loop time by the number of performed loops.

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.