data.table.threads v1.0.1
New Features
- Added a function that computes and adds recommended efficiency plot
data (speedup lines and thread count points) based on the
recommendedEfficiency argument (has been removed from
findOptimalThreadCount). This allows the user to construct
plots without recommended efficiency lines and points, and on the other
hand, to easily add them when needed with the help of this
function.
findOptimalThreadCount now has
recommendedEfficiency, an argument for adjusting the slope
of the recommended speedup trend/line and the recommended thread count
designating point in speedup plots generated by using the plot method on
an object returned by the function.- The data used for plotting (consisting of speedup trends/lines and
key points) is now accessible via attributes of the
data.table returned by
findOptimalThreadCount.
- Added new arguments
verbose and times to
runBenchmarks and findOptimalThreadCount,
allowing users to enable/disable detailed message printing and specify
the number of benchmark repetitions.
- S3
plot method constructs a more intuitive
ggplot for data_table_threads_benchmark class
objects.
- Only user-facing functions are exported.
- Examples have been updated to use namespace qualification.
- Added a function to set the thread count for
data.table
operations based on recommended/optimal performance (in benchmarks) of a
user-specified data.table function.
Improvements
- Extensively refactored my codebase to remove redundancy.
- Made several optimizations (using
data.table over
data.frame, implementation of faster data aggregation
operations, minimizing overhead for various function calls, etc.).