Project

General

Profile

Cluster tests Nov 12 2012

To capture a snapshot of the cluster at LCC before it is shipped to LNGS, I'll run some artdaq tests with multiple hosts and generated data.

For this test, I checked out the latest artdaq code, modified the EventStore class to print out its statistics at the end of the job, and built artdaq.

These tests used commands like this one:
  • daqrate 1 1 10 101 --nodes=dsfr3,dseb3 --ddnodes=1 -- -c daqrate_gen_test.fcl
  • where FHICL_FILE_PATH has been modified to include my home directory and daqrate_gen_test.fcl has the contents listed below.
Results:
  • 1 1 10 101 --nodes=dsfr1,dseb1 --ddnodes=1 - 1.0 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr2,dseb2 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr3,dseb3 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr4,dseb4 --ddnodes=1 - 1.0 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr4,dseb5 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr4,dseb6 --ddnodes=1 - 1.0 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr1,dsfr2 --ddnodes=1 - 1.4 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr1,dsfr3 --ddnodes=1 - 1.4 GB/sec per EventStore
  • 1 1 10 101 --nodes=dsfr1,dsfr4 --ddnodes=1 - 1.3 GB/sec per EventStore
  • 1 1 10 101 --nodes=dseb1,dseb2 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 1 1 10 101 --nodes=dseb1,dseb3 --ddnodes=1 - 1.2 GB/sec per EventStore
  • 1 1 10 101 --nodes=dseb1,dseb4 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 1 1 10 101 --nodes=dseb1,dseb5 --ddnodes=1 - 1.2 GB/sec per EventStore
  • 1 1 10 101 --nodes=dseb1,dseb6 --ddnodes=1 - 1.1 GB/sec per EventStore
  • 2 2 10 101 --nodes=dsfr{1-2},dseb{1-2} --ddnodes=2 - 1.2 GB/sec per EventStore
  • 2 2 10 101 --nodes=dsfr{3-4},dseb{3-4} --ddnodes=2 - 1.2 GB/sec per EventStore
  • 2 2 10 101 --nodes=dsfr{3-4},dseb{5-6} --ddnodes=2 - 1.2 GB/sec per EventStore
  • 2 2 10 101 --nodes=dsfr{1-4} --ddnodes=2 - 1.3 GB/sec per EventStore
  • 2 2 10 101 --nodes=dseb{1-4} --ddnodes=2 - 1.3 GB/sec per EventStore
  • 2 2 10 101 --nodes=dseb{3-6} --ddnodes=2 - 1.3 GB/sec per EventStore
[biery@dsfr3 artdaqBuild]$ cat /home/biery/daqrate_gen_test.fcl
BEGIN_PROLOG
payload_size: 524288
det:
{
  generator: GenericFragmentSimulator
  payload_size: @local::payload_size
  run_number: 101
  fragments_per_event: 1
  events_to_generate: 5000
}
END_PROLOG

daq:
{
  use_art: false
  max_payload_size: @local::payload_size
  detectors: [ det0, det1, det2, det3, det4 ]
  fragments_per_source: 5000
  det0: @local::det
  det1: @local::det
  det2: @local::det
  det3: @local::det
  det4: @local::det
}

det0.starting_fragment_id: 0
det1.starting_fragment_id: 1
det2.starting_fragment_id: 2
det3.starting_fragment_id: 3
det4.starting_fragment_id: 4