# Understanding Magma issue with maximal subgroups computation

I am trying to compute the maximal subgroups of the wreath product $$(\mathbb Z/10\mathbb Z)\wr S_{99}$$ using Magma's algorithm for maximal subgroups, which is an implementation of an algorithm of Cannon and Holt. The computation is not completed, presumably due to the large order of this group. The only output I get is 'Killed: 9'. I would like to know how to interpret this. Is it an issue with memory? If so, what part(s) of the computation might be causing this?

I just ran this calculation on a machine with lots of memory, and it completed in just over two hours using about 85GB of memory. There are 59 classes of maximal subgroups.

You can follow the progress of the computation by turning on verbosity. For this one I would recommend

SetVerbose("Subgroups",3);

The calculation proceeds by finding the maximal subgroups of $$S_{99}$$, which is quick and easy, and then lifting them through elementary abelian layers of sizes $$2^1$$, $$2^{98}$$, $$5^{98}$$, $$5^1$$.

The lifts through the elementary abelian $$2$$-groups are not so difficult, and complete using about 8GB at most. The difficult one is $$5^{98}$$. This involves some cohomological calculations that need a lot of memory. Unfortunately there are some aspects of the implementation that are profligate with memory in examples of this type. For example some of the intermediate subgroups that arise in the calculations have unnecessarily large numbers of generators, which result in an excessive number of $$98 \times 98$$ matrices being calculated.

Anyway, the crash that happened for you was almost certainly due to running out of memory.

Probably your program received a "kill" signal (they are numbered, SIGKILL is number 9), typically these signals are sent by programs run by your OS, "OOM-killers" (out of memory). TL;DNR, you use too much memory.