Is it possible to test intel speed shift technology with CPU benchmark scores?

[karry]

I want to test whether intel speed shift technology related to CPU in BIOS is effective, so I want to test CPU Benchmark by running Passmark performance in the system, and find that the difference between speed shift and speed Shift is more than 40%. Does this show that this technology has taken effect?
ps: The power plan on the OS is powersave mode

[karry]

I see that the following option in BIOS will affect P-states, may I ask whether this option will bring the above performance difference?

[David (PassMark)]

If this is the only thing you changed between the two runs, and the results are repeatable, then I think there is only one conclusion.

But it might be that the O/S doesn't take over the power management role if the CPU is capable of controlling it's own frequency, as it would on older CPUs.

[karry]

Can I confirm whether this technology (intel speed shift) is effective through performance testing?

[David (PassMark)]

You just did, didn't you?

[karry]

Yes, my test results are so, but I am not sure whether this performance difference is caused by the Power plan selecting powersave under OS?

However, we only test the BIOS for this change, and the OS is all set to powersave mode.

I conducted tests on different Intel platforms: the 6th generation skylake, the 8th generation coffee lake and the 13th generation RPL, and the test results are consistent. If Intel Speed Shift technology is enabled, the CPU performance will be improved by 40%~50% compared with that if it is not enabled

[David (PassMark)]

I have never looked into the detail of speed step, but I assume if it is disabled, then the CPU will be running a lower clock speeds when under load.

[karry]

Intel early proposed a dynamic voltage frequency switching technology SpeedStep, with the operating system, for different workloads to allocate different performance and power state, but the disadvantage is that the state switching speed is slow, the fastest is 30 milliseconds, the system can control the state is limited.
Speed Shift bypasses the operating system, the processor communicates directly with the power control unit, and opens all power states to the operating system, which can complete the state switch in as little as 1 millisecond. As shown below:
So I think the performance is definitely high after the turn on（Speed Shift）, but once the turn off, the CPU benchmark score will not be affected by the power plan(Powersave mode) under the system, so that the test result is very low.