Performance Default Overrides

  -proc     Number of CPU processors used
  -cons     Ratio of parsers to processors
  -serv     Concurrent parser instances
  -chan     Communication channel depth
  -heap     Order restoration heap size
  -farm     Node allocation buffer length
  -gogc     Garbage collection tuning knob

Internal Component Performance

  -chunk    StreamBlocks
  -split    StreamBlocks -> SplitPattern
  -token    StreamBlocks -> StreamTokens

Debugging

  -debug    Display run-time parameter summary
  -empty    Flag records with no output
  -ident    Print record index numbers
  -stats    Show processing time for each record
  -timer    Report processing duration and rate
  -trial    Optimize -proc value, requires -input

Record Set Indexing

  -index    Place next record size tag before object
  -turbo    Use <NEXT_RECORD_SIZE> values for speed

Documentation

  -keys     Keyboard navigation shortcuts
  -unix     Common Unix commands

Performance Tuning Script

  XtractTrials() {
    echo -e "<Trials>"
    for tries in {1..5}
    do
      xtract -debug -input "$1" -proc "$2" -pattern PubmedArticle -element LastName
    done
    echo -e "</Trials>"
  }

  for proc in {1..8}
  do
    XtractTrials "carotene.xml" "$proc" |
    xtract -pattern Trials -lbl "$proc" -avg Rate -dev Rate
  done

Processor Titration Results

  1    27622    31
  2    51799    312
  3    74853    593
  4    95867    1337
  5    97171    4019
  6    93460    2458
  7    87467    1030
  8    82448    2651

Entrez Index Performance Measurement

  IndexTrials() {
    echo -e "<Trials>"
    for tries in {1..5}
    do
      cat "$1" | xtract -debug -proc "$2" -e2index
    done
    echo -e "</Trials>"
  }

  for proc in {1..8}
  do
    IndexTrials "carotene.xml" "$proc" |
    xtract -pattern Trials -lbl "$proc" -avg Rate -dev Rate
  done

MeSH Tree Index Preparation

  nquire -dwn ftp://nlmpubs.nlm.nih.gov online/mesh/MESH_FILES/xmlmesh desc2021.zip
  unzip desc2021.zip
  rm desc2021.zip

  cat desc2021.xml |
  xtract -pattern DescriptorRecord -element "DescriptorRecord/DescriptorUI" \
    -sep "," -element TreeNumber > meshtree.txt

Execution Profiling

  cat carotene.xml > /dev/null
  ./xtract -profile -timer -input carotene.xml -pattern PubmedArticle -element LastName > /dev/null
  go tool pprof --pdf ./cpu.pprof > ~/Desktop/callgraph.pdf
  rm cpu.pprof