Some DTC benchmarks are missing from the webpage.

[p16i]

According to config.yaml, mlpack and shogun seem to have configuration for the DTC task. However, the benchmark webpage seems to have only scikit's results.

Is it possible that the current data is outdated? How can we update the newest results?

[zoq]

You are right, the DTC benchmark was added and the results haven't been updated yet. I'll see if I can start the benchmarks in the next day or two.

[p16i]

thanks!

[zoq]

Okay, I've started the shogun benchmark job (http://ci.mlpack.org/view/benchmarking/job/benchmark%20-%20run%20-%20shogun/3/) hopefully it does finish without any errors.

[p16i]

looks like we have some errors from the build. From the build log, I can't find that much information. It just says:

Makefile:180: recipe for target '.run' failed
make: *** [.run] Error 137
Build step 'Execute shell' marked build as failure
Finished: FAILURE

@zoq Do you have more detailed information?

[zoq]

I think the error message got cut off, because the print was buffered, I'll set export PYTHONUNBUFFERED=true to dump log messages immediately and restart the benchmark.

[zoq]

Restarted: http://ci.mlpack.org/view/benchmarking/job/benchmark%20-%20run%20-%20shogun/7/