Benchmark tools now accept options such as 'nb-nodes' and 'seed' instead
of default hard-coded values.
Details
Diff Detail
- Repository
- rDGRPH Compressed graph representation
- Lint
Automatic diff as part of commit; lint not applicable. - Unit
Automatic diff as part of commit; unit tests not applicable.
Event Timeline
Build has FAILED
Link to build: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/8/
See console output for more information: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/8/console
java/server/src/main/java/org/softwareheritage/graph/benchmark/Common.java | ||
---|---|---|
53–54 | I'm assuming 0 here is the default seed value that will be used if it's not given on the command line. The default for random seed should not be zero (or anything fixed), rather it should be something backed by entropy. As I understand from the java documentation, the way to obtain that is simply not passing an argument to the Random constructor. So you should change this to make sure that, if "seed" is not passed on the CLI, no constructor is passed to Random. |
Build has FAILED
Link to build: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/9/
See console output for more information: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/9/console
java/server/src/main/java/org/softwareheritage/graph/benchmark/AccessEdge.java | ||
---|---|---|
32 | Cheating ! ;-) You should make the default value for the seed null (it seems to be valid in JSAP) and check whether that value is null as condition here. |
java/server/src/main/java/org/softwareheritage/graph/benchmark/AccessEdge.java | ||
---|---|---|
32 | The problem with null is that in Java only wrapper class for data types can be checked against null, so in this case Long instead of long which I found a bit weird compared to the -1 "trick" :p |
Build has FAILED
Link to build: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/10/
See console output for more information: https://jenkins.softwareheritage.org/job/DGRPH/job/tox/10/console
java/server/src/main/java/org/softwareheritage/graph/benchmark/AccessEdge.java | ||
---|---|---|
32 | You can either use a Long benchArgs.seed attribute and use getLong() later, or, have a benchArgs.rng instead of .seed and initialize it earlier on instead of using config.getLong. Or am I missing something? |