If you emphasis on reading comprehension, the actual questions was how many processor cores is need to verify the incoming transactions + the mined block?
Since you asked, I just looked in the log file on a 4-year old Xeon server I'm using as a node. It takes approximately 0.14 seconds from the time a new transaction arrives until the time it is relayed. I believe most if not all of this processing is single threaded, which suggests approximately 7 transactions per second per core on a 4-year old CPU.
On the order-of-magnitude of 20 txs/sec per core on a late model CPU, i.e. an order-of-magnitude higher than Bitcoin (<1 tx/s) now per core but two orders-of-magnitude less than Visa (2 - 6K tx/s) now per core, means Monero (Cryptonote) can't scale to any where even close to global Visa scale and remain both decentralized for mining with fast block period (thus fast transactions), not to mention the likely order(s)-of-magnitude more scaling above that to reach ubiquitous global micro transactions and programmable contracts on the block chain.
So you would have to solve both this and the blockchain bloat in order to scale to global widespread use. It appears that one-time ring signatures are fundamentally incompatible with scaling.
Cryptonote can't encourage too much use with zero transaction fees, because it can't accept the scaling that can come with it.
I believe Zerocash has similar scaling issues. DarkCoin (and CoinJoin) has the simultaneity problem that fights scaling because to mix you need someone else who wants to mix with you at the same denominations at the same time (not mention being either theoretically defeated with jamming and/or Sybil attack on masternodes) and to perform this meeting with scaling you need global coherence on submitted txs which means either centralization (synchronicity) or no scaling.