Being within 2 -3 times slower than the top of the line is quite reasonable for a new chip. The Java results being so much slower are likely the difference in

Being within 2 -3 times slower than the top of the line is quite reasonable for a new chip. The Java results being so much slower are likely the difference in optimizing for new arch, bot neseccarily the chip itself. Given that many apps lice quite nicely on VM and containers (using a lot less capacity then is available from the hardware), I think this is respectable to start with.