DDN IME Evaluation
Go to file
Eugen Betke 67ae1ba065 Lustre stryping info 2018-10-24 11:19:31 +02:00
info Lustre stryping info 2018-10-24 11:19:31 +02:00
output NN:8 experiment results 2018-10-24 07:42:09 +02:00
output.old Results with 128GB test file 2018-10-21 22:38:23 +02:00
output_v2 Striping benchmark (read only) 2018-10-24 11:10:10 +02:00
LICENSE Initial commit 2018-10-04 16:02:57 +02:00
README.md Update README.md 2018-10-09 11:13:31 +02:00
batch_job.sh Comments 2018-10-09 10:49:40 +02:00
drop_caches.sh Drop caches script 2018-10-21 20:46:34 +02:00
eval_analysis.R Analysis and makeDB script fixed 2018-10-23 10:25:28 +02:00
eval_netcdfbench.R Evaluation scripts 2018-10-04 16:53:44 +02:00
eval_openclose.R Evaluation scripts 2018-10-04 16:53:44 +02:00
eval_performance.R Evaluation scripts 2018-10-04 16:53:44 +02:00
eval_readwrite.R Evaluation scripts 2018-10-04 16:53:44 +02:00
micro_run.py Minor script changes 2018-10-24 11:12:22 +02:00
mkdb.py Analysis and makeDB script fixed 2018-10-23 10:25:28 +02:00
run.sh Minor script changes 2018-10-24 11:12:22 +02:00
run_v2.sh More advanced benchmark script (v2) 2018-10-24 11:12:52 +02:00
tool_estimate_time_remains.sh Tool for estimating runtime 2018-10-24 08:05:59 +02:00
tool_sanity_check.sh Parameters check for IOR 2018-10-23 10:23:35 +02:00
tool_show_bad_output_files.sh Minor script changes 2018-10-24 11:12:22 +02:00
tool_show_good_nodes.sh New name 2018-10-23 10:24:50 +02:00

README.md

I/O-Performance exploration

Running performance tests

sbatch ./batch_job.sh

Put all results in a SQLite3 data base

./mkdb.py <perf_results_folder> <db_file_name.db>

Evaluation with R

  1. ./eval_analysis.R <db_file_name.db> <output_folder>
  2. ./eval_netcdfbench.R <db_file_name.db> <output_folder>
  3. ./eval_openclose.R <db_file_name.db> <output_folder>
  4. ./eval_performance.R <db_file_name.db> <output_folder>
  5. ./eval_readwrite.R <db_file_name.db> <output_folder>