ndaniels / mrfy-bench Goto Github PK

README for `viterbi-bench`
==========================
See `viterbi-bench -h`


README for `mrfy-bench`
=======================
This is a simple benchmarking framework for MRFy.
It is not yet intended for public consumption.
There are four basic variants: `mrfy-bench <-s | -m | -l | -p>`
Short is intended as a microbenchmark; it takes 5-10 seconds on reasonable 
hardware.
Medium takes less than 2 minutes, and takes the better of two passes for each.
Long and Paper take a significant amount of time, running more templates and
choosing the best of 5 runs for each.
The only difference between Long and Paper is that the latter requires SMURF
to be run as well; Long only runs SMURF if either the SMURF_DAT_PATH environment
variable is set, or the path is passed as an argument after `-p`.
Medium runs SMURF under the same criteria as Long.
Note: mrfy and smurf (if it is being run) must be in your path.