I'd check if tcpdump is the first bottleneck. It has to do a lot of work and I doubt it is multithreaded. "time tcpdump -options >/dev/null" gives the minimum run time for the perl script. Look at the load or "top" while doing this.

Recording smaller tcpdump_infiles or splitting them up at a good boundary allows running multiple tcpdump processes.

OTOH it might be justified to parse the tcpdump_infile directly from perl.