Store intermediate de Bruijn data in tmp folder rather than in RAM

[yqy6611]

Is your feature request related to a problem? Please describe. For generic questions use Q&A section in the Discussions forum above.

SPAdes (especially metaSPAdes) requests too much memory. Exploring temporary files generated, only k-mer data is stored. Intermediate de Bruijn data seemed to be stored and manipulated in memory. For large scale metagenomics, the requested memory is impractical to be satisfied.

Describe the solution you'd like

Rather than in RAM, intermediate de Bruijn data should be stored in tmp folder assigned as well. Truly it will reduce the speed, yet such feature will make large-scale co-assembly practical.

Describe alternatives you've considered

No response

Additional context

No response

[asl]

Thanks for your suggestions. Patches to improve SPAdes performance are always welcome!