[NMF-user] nmf error

[Renaud Gaujoux]

> I have produced a matrix of a similar size and am still encountering the
> same problem- it will only run when nrun=1.

This does not break on my side.

> I have attached the log file with .options='d' and also the log file running
> it without this option.

The log file with option 'd' does not show any error, does it?
The other does (NB: you can use traceback(,10) to limit the output),
which is strange since it should use the same seed.
>From the log, it appears to break when computing the final objective
value (after convergence).

If it also breaks with `nrun=1`, then let's use this value for
debugging, as it discards parallel computing issues.
Can you please try adding argument `copy=TRUE`?

Thanks