Suppose you have a system that under some workload will draw 10W to run the chipset, WiFi, SSD and so on and another 15W to run the CPU, so that it's 25W in total. The 15W CPU has X number of cores and gets a score of 1000 on some benchmark. Performance per watt is 40 points per watt.
If you put 4X of the same cores in the same system then the CPU uses 60W, the WiFi etc. still use 10W, so the whole system uses 70W. It scores 4000 points and the performance per watt is 57 points per watt. You got four times the multi-thread performance and performance per watt went up.
alt Hacker News
Suppose you have a system that under some workload will draw 10W to run the chipset, WiFi, SSD and so on and another 15W to run the CPU, so that it's 25W in total. The 15W CPU has X number of cores and gets a score of 1000 on some benchmark. Performance per watt is 40 points per watt.
If you put 4X of the same cores in the same system then the CPU uses 60W, the WiFi etc. still use 10W, so the whole system uses 70W. It scores 4000 points and the performance per watt is 57 points per watt. You got four times the multi-thread performance and performance per watt went up.