Tjonnie, Steve P, Les, Chad, Jolien, FIXME.
Optimization efforts: gstlal_inspiral_flopulater and function call graphs
About gstlal_inspiral_flopulator:
1) Les, analyzing 50Msun to 350Msun, 100 templates per bank, using non-optimized libraries
flopulater results: https://ldas-jobs.phys.uwm.edu/~lwade/Search/optimization_tests/IMBHB_templates/flopulate.txt
function call profiling: https://ldas-jobs.phys.uwm.edu/~lwade/Search/optimization_tests/IMBHB_templates/profile.txt
2) Tjonnie, running with 200 templates per sub bank, lower mass range, down to mchirp = 5Msun, and non-optimized libraries
flopulator: https://ldas-jobs.phys.uwm.edu/~tgfli/allflops.txt
function call profiling: https://ldas-jobs.phys.uwm.edu/~tgfli/profile.txt
In both cases, audioresample dominates the computational cost. BLAS is another high contender for computational cost, but subject to improvement after linking against optimized libraries installed at UWM.
Code issues: