I understand that short power PL2 allows the processor to run at higher watt. I also understand what clamp means.
However, what does the turbo time limit exactly mean? Is there a "time guard space" between how often a turbo can occur? How often can a CPU go turbo mode per timeslot?
I have done a benchmarking experiment because I didn't found much information about a possible time guard space. PL1=15[W], PL2=25[W], time=10. The CPU went full usage mode with prime95, that is it used the most possible wattage. Before it was in idle.
My observations were the following: When the CPU hits the PL1 limit from below, that is power consumption goes up from 1[W] to PL1[W], it gets a turbo time space for 10. That happens right after I started the benchmark. Afterwards, the PL1 limit is applied again. If the power consumption is still as high as PL1, which it is because of the stress test, processor won't get another turbo. It only gets another PL2 turbo if the wattage goes down significantly below the PL1 limit, for example if I stop and restart the benchmark.
My conclusion is that the PL1 and PL2 limits work like a hysteresis. You reach the PL1 limit from below, you get a turbo. You reach the PL1 limit from above, that is you just were in a PL2 turbo and the CPU still needs to be throttled by PL1, you won't get another turbo. You need to go down way lower than the PL1 limit until the PL1 doesn't throttle anymore. Then you get another PL2 turbo. There doesn't seem to be any time guard space between two PL2 turbos though, like a "a turbo can only occur after the previous one ended for 5 seconds".
Does this make sense?
However, what does the turbo time limit exactly mean? Is there a "time guard space" between how often a turbo can occur? How often can a CPU go turbo mode per timeslot?
I have done a benchmarking experiment because I didn't found much information about a possible time guard space. PL1=15[W], PL2=25[W], time=10. The CPU went full usage mode with prime95, that is it used the most possible wattage. Before it was in idle.
My observations were the following: When the CPU hits the PL1 limit from below, that is power consumption goes up from 1[W] to PL1[W], it gets a turbo time space for 10. That happens right after I started the benchmark. Afterwards, the PL1 limit is applied again. If the power consumption is still as high as PL1, which it is because of the stress test, processor won't get another turbo. It only gets another PL2 turbo if the wattage goes down significantly below the PL1 limit, for example if I stop and restart the benchmark.
My conclusion is that the PL1 and PL2 limits work like a hysteresis. You reach the PL1 limit from below, you get a turbo. You reach the PL1 limit from above, that is you just were in a PL2 turbo and the CPU still needs to be throttled by PL1, you won't get another turbo. You need to go down way lower than the PL1 limit until the PL1 doesn't throttle anymore. Then you get another PL2 turbo. There doesn't seem to be any time guard space between two PL2 turbos though, like a "a turbo can only occur after the previous one ended for 5 seconds".
Does this make sense?