Thanks for the update. I had the impression that limit is only on the cli-wallet side.
BTW, the testnet has reached 2,753 txs/block or 917.67 txs/sec already - tho it could be that the blocktime was 5 seconds back at that time. Need to check. At 5 sec we would still be at 550tps which is still "good".
We are looking into the p2p code and will have to deal with the results of this first stress test in subsequent tests.