Permalink
Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
107 lines (71 sloc) 2.07 KB
==== Run
SIZES="50k 250k" runPerf
Tghsi generates the report into Reports/
==== Process
Offical results:
http://wifo5-03.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/results/V7/index.html#detailsExploreCompare
See run-complete for a script to drive all the scripts.
mkdir Data
mkdir TDB-DB
genAll (choose sizes wanted)
SIZES="50k 250k" loadAll (choose sizes to load)
SIZES="50k 250k" runPerf (choose sizes to test)
# generate results
runResults
Key results file:
bsbm_query_and_size_t1ables_of_stores.html
--> bsbm-query-results.html
Or:
Collection of dated runs under Runs/
genAllResults
Also:
DATE=[date 2010-MM-DD]
mkdir Runs/$DATE
mv Results/* Runs/$DATE
# Note: the Java code knows the range of possible store sizes
# If you miss 200m out, say, you get
# "Error: XML result file names must contain size metrics e.g. store_50k.xml for res-200m.xml"
# Run results and create bsbm-results.html in the Runs/DATE/ directories:
genAllResults
== Layout:
Data/data-SIZE/...files...
Results/res....
genAll creates all the datafiles laid out like this.
== BSBM details
The original benchmark uses the web API and ...
# Execute single-client test run (500 mixes performance measurement, randomizer seed: 808080)
# Execute multiple-client test runs. ( 4 clients, 500 query mixes, randomizer seed: 863528)
# Execute test run with reduced query mix. (repeat steps 2 to 4 with reduced query mix and different randomizer seed 919191)
== Generate
mkdir Data
# Alter sizes
genAll
== Query
(can set SIZE in "db-size")
SIZES="50k 250k" loadAll (choose sizes to load)
SIZES="50k 250k" runPerf (choose sizes to test)
== Run Generator
50k: 128
250k: 666
5m: 14088
1M: 2785
25M: 70812
100M: 284826
200M: 570000
== Results
java -cp bin:lib/* benchmark.tools.ResultTransform ??????
Data layout:
StoreName/25m.xml etc etc
== Setup
??
TestDriver reads "-ucf FILE"
Default is userses/explore/sparql.txt
Use case file is lines of
querymix=rel filename
Here, set to:
-ucf usecase.txt
==>
querymix=DIRECTORY
querymix=
for current directory
Unreliable