Skip to content
aalbrechtsen edited this page Sep 7, 2026 · 1 revision

For large sample sizes and if many threads are allocated to the program the memory usage can become severe.

The filewriting has to be done in the correct order, and if some given region of the genome is computational intensive. You might experience an increase of memory usage, due to the fact that the "done" chunks are waiting to be written. To avoid this problem you can specify a maxsize of queue of "done" chunks.

-nQueueSize

Clone this wiki locally