> The 90 minutes is not - the blockchain has grown quite a lot since last
> year, and as for the 3.5 speed, I havn't tested it since Pieter's
> ultraprune - libcoin also has something similar to ultraprune, done
> directly in the sqlite database backend, but I should run a head to head
> again - could be fun. I would assume, though, that the result would be
> similar timings.
>

ultraprune made a huge difference. I think it's very likely that this claim
is no longer true. Bitcoin got a lot more optimised since you first did
libcoin.