razzo_project
| Branch |  |
|---|---|
master |
 |
develop |
 |
richel |
 |
razzo experiment scripts and storage.
The research project uses these GitHub repo's:
- razzo: R code
- raztr:
razzoresults of a test run - razzo_project: bash scripts to run and analyse an experiment
- razzo_article (private): scientific manuscript (private GitHub for now)
- razzo_pilot_results (private): results of the pilot runs
Versions
- nLTT v1.4.3
- beautier v2.3.2
- tracerer v2.0.2
- beastier v2.1.1
- mauricer v2.0.5
- babette v2.1.1
- mcbette v1.8
- pirouette v1.6.2
- nodeSub v0.4.2
- mbd v1.0
- mbd.SimTrees v0.1
- becosys v1.0.1
- peregrine v1.0
- razzo v0.7.2
Code can be found in the packages folder.
Data
To create the table in overview.md:
cd razzo_project
./scripts/90_collect_run_times.shSee overview.md.
20190908is the favorite starting point- The
n_particlesin20190908and20190910is correct to the actual value of 1, due to amcbetteerror
To download the data,
download http://richelbilderbeek.nl/razzo_project_[date].zip where [date] is the date,
for example: http://richelbilderbeek.nl/razzo_project_20190801.zip
Our verdicts if a setting is good enough: see verdict.md.
Note:
- Bug #350 had its unfluence up until #346. #347 is without that bug
Folder structure
razzo_project has the following folder structure:
scripts: contains the scripts
The scripts in scripts create the following extra folders:
data: contains the simulation dataresults: contains the simulation results
Cluster usage
:warning: must run from the root folder
Regenerate testing data:
sbatch ./scripts/regen_data.sh
Run the razzo experiment:
sbatch ./scripts/1_install_razzo.shsbatch ./scripts/2_create_parameter_files.sh testorsbatch ./scripts/2_create_parameter_files.sh fulsbatch ./scripts/3_run_razzo.shsbatch ./scripts/7_create_nltt_stats_file.shsbatch ./scripts/8_create_esses_files.shsbatch ./scripts/9_create_marg_liks_file.shsbatch ./scripts/10_create_fig_1_file.shsbatch ./scripts/11_create_n_taxa_file.sh
Local usage
Same, but without sbatch.
See .travis.yml for the complete usage.
mcbette timeseries
| n_taxa | n_nucleotides | run_time |
|---|---|---|
| 10 | 1k | 0:40 or 40 mins |
| 20 | 1k | 1:23 or 83 |
| 40 | 1k | 3:35 or 215 mins |
| 80 | 1k | more than 10 hours, got cancelled |