It took about 10 hours to compute the consolidation graph (including time consuming blast computations) and only several minutes to perform the refinement procedure.
In comparison, a larger 100 × 100 × 100 grid would take an impossible three million CPU hours to compute (equivalent to 343 years running on a single computer).
It takes several hours to compute a model this complex – and building the software that can generate these models takes significantly longer still.
Even when using GPUs, however, it takes hours to compute full three-dimensional iterative reconstructions for typical data sizes in modern large-scale tomographic experiments [18].
Resulting maps were similar to the map presented, but required more than 24 hours to compute.
We also found that it took around 9 hours to compute the closure and the consolidation graph assuming precomputed blasts.
Our implementation of the canonical basis approach, on the other hand, takes within minutes to hours to compute a few thousand EMs.
The performance scales roughly linearly with the amount of DNA to be processed; for example, a medium sized metagenome (C1-oxidisers in lake water [ 48], at 37 Mb) requires about 30 hours to compute on a single CPU; a larger metagenome (220 Mb from a hot spring) requires close to 200 hours.
Modern, high-level calculations can take hours or days to compute using fast computers.
