Crash on ensembling classification data of big size

"Ensemble Classification Data" workflow element is not able to process big data. A crash occurs. See, for example, the scenario below.

Scenario:

1. Open "Parallel NGS reads classification" workflow sample, set "Single-end" reads.
2. Select "Read NGS SE Reads" element, input file "C_S10_final_R1.fastq" (it is about 5.9 Gb).
3. Optionally, remove the MetaPhlAn2 element as it will not affect the result.
4. Run the workflow.
   Current result: all classification steps (CLARK, Kraken and DIAMOND) pass. The ensemble element start working and the application crashes.
   A CSV file created by the ensemble element is empty.

The same result happens in case of paired-end version of the workflow when only two classifiers are present (CLARK and Kraken).