I noticed sudden spikes of the CPU load from the testnet Bitcoin Core instances. High and prolonged enough to turn on the fans in my machines, which is very rare. Unfortunately I don't have time to investigate this myselves. I see nothing out of ordinary in the debug.log, ordinary for the testnet that is. mempool is also somewhat high, again for the testnet.
AFAIK BIP 101 supporters split the testnet. They created a high number of empty blocks. After the split there was a huge fork ~12 hours, difficulty is now reset (1.000) and the usual chaos after a diff reset happens.
At least thats what I picked up durring the day, mainly from here ->
https://www.blocktrail.com/tBTC/blocks/1and a thread
I cant find atm... here ->
https://bitcointalk.org/index.php?topic=1241662.0Thanks for your suggestions.
Well, difficulty reset to 1.000 happens all the time since Gavin made that change years ago; when the most recent block is more that 20(?) minutes old. Long forks also happen quite often on testnet, but I don't recall such a CPU-intensive one. I will have to put more computers online with BIP101 compiled in.