My bad. I was under the impression that Intel only made single-core Celerons with hyper-threading disables, after I saw a couple of them in NAS's and ones running FreeDOS.

I always had this impression that the modern Pentiums they make were dual-core or maybe sometimes quad-core, as a lower tier to the Core i3 lineup.

I have a server running a Xeon E5-2650 which is at 2GHz and that also runs Core kinda slow, not really abysmal as OP's waiting times but not as fast as I'd like, which is speed of the node running on my E3-1240 at 3.3GHz.

---

Still though, the L1 cache on that J1900 is really small at 2MB, my 2650 has 30MB by comparison. Given that Bitcoin Core is written in C++, I'd love to see some optimization that pushes more of the transaction that's currently being verified on the L1 cache which is much, much faster than main RAM.