There were so many claims in the first couple of years after Perl6/Raku's initial release in December 2015 about how it had the potential to be optimised due to its superior design but here we are nearly 8 years later and it's still takes 10 times as long as Perl 5 to parse a log file with a regex.
Yes, and those claims might have even been real (also, Perl still lags behind on Unicode stuff, or so I heard?)... the thing we like much less to talk about is that there just needs to be somebody to make those optimizations happen. Work on the actual bytecode VMs (or rather just MoarVM at this point) has been desperately lacking.
Jonathan Worthington has been much going on and off in the recent years. He still had one big optimization project that got completed in 2021 - the new dispatch mechanism - but I don't know about the benchmarks and I doubt it would help with the shear throughput of basic data processing.
