Skip to content

Latest commit

 

History

History
45 lines (38 loc) · 1.64 KB

Energy-Measure.md

File metadata and controls

45 lines (38 loc) · 1.64 KB
Raw

Conducting an energy measure

Start processing on jetson:
sudo python efficient-gpu-joins/impl/benchmark/zmq_bench.py [raspberry ip] [settings file] 0

Start measure on raspberry pi:
python efficient-gpu-joins/impl/benchmark/zmq_measure.py

Results

Measurement results are stored on the jetson in:
python efficient-gpu-joins/impl/measures/[current timestamp]

Results given by stream join executable: [setting]bench.csv Results given by power measurement executable: exp[setting]

Settings

Setting files store configurations used to start an experiment. Each line represents one experiment.
efficient-gpu-joins/impl/benchmark/settings/

Usage:
  -n NUM   number of tuples to generate for stream R
  -N NUM   number of tuples to generate for stream S
  -O FILE  file name to write result stream to
  -r RATE  tuple rate for stream R (in tuples/sec)
  -R RATE  tuple rate for stream S (in tuples/sec)
  -w SIZE  window size for stream R (in seconds)
  -W SIZE  window size for stream S (in seconds)
  -p [cpu, gpu]  processing mode (cpu or gpu)
  -s SEC  idle window time
  -S SEC  process window time
  -T enable sleep time window
  -t sleep control in worker
  -b NUM  batchsize for stream R
  -B NUM  batchsize for stream S
  -g NUM  GPU gridsize
  -G NUM  GPU blocksize
  -f enable frequency by stream join
  -e end when worker ends

Postprocessing

Postprocessing includes generating energy plots and calculating the average energy consumption.
Running postprocessing scripts:
efficient-gpu-joins/impl/postprocess/postprocess.sh efficient-gpu-joins/impl/measures/[timestamp]