Users frequently want to know the approximate runtime of algorithms so they can plan, e.g. computational use, queue size, etc.
On the algorithm side, the only parameters that alter runtime for gibbs are total number of passes through gibbs_sampler, number of sinks, and total number of sequences in the sinks.
On the hardware side, gibbs is CPU rather than memory intensive, so processor speed is the most important parameter for determining runtime.
We could provide a simple method that just tests a grid of: number of sinks X total sequences X total passes X number of cores and returns the user a very simple plot that would allow them to estimate run time.
Users frequently want to know the approximate runtime of algorithms so they can plan, e.g. computational use, queue size, etc.
On the algorithm side, the only parameters that alter runtime for
gibbs
are total number of passes throughgibbs_sampler
, number of sinks, and total number of sequences in the sinks.On the hardware side,
gibbs
is CPU rather than memory intensive, so processor speed is the most important parameter for determining runtime.We could provide a simple method that just tests a grid of: number of sinks X total sequences X total passes X number of cores and returns the user a very simple plot that would allow them to estimate run time.