Optimal CPU Tuning settings for i9 MBP to stop VRM throttling / Explanation of Apple's Engineering F

[f54da]

>So, Apple has shipped the i9 Macbook Pro to 125 Watt Turbo for 28 seconds, 100 Watt Steady state.

@winterny So this is interesting, I just checked my 2014 MBP which is specced for ~50W TDP, and I see the exact same thing reading from the power MSR: 100W PL1, 125W PL2. You can see searching the comments on the https://github.com/sicreative/VoltageShift/issues that this is true for at least the 2014 and 2015 MBPs. This makes things interesting, why does Apple not bother setting the power limit MSR to the proper values, and why was this less noticeable on the 2013-2015 MBPs?

 

[leman]

>So, Apple has shipped the i9 Macbook Pro to 125 Watt Turbo for 28 seconds, 100 Watt Steady state.

@winterny So this is interesting, I just checked my 2014 MBP which is specced for ~50W TDP, and I see the exact same thing reading from the power MSR: 100W PL1, 125W PL2. You can see searching the comments on the https://github.com/sicreative/VoltageShift/issues that this is true for at least the 2014 and 2015 MBPs. This makes things interesting, why does Apple not bother setting the power limit MSR to the proper values, and why was this less noticeable on the 2013-2015 MBPs?

Apple did not restrict the TDP on Intel Macs. They throttle on temperature instead. This has the benefit of getting better performance if you still have thermal headroom available. For example, assume your chassis and ambient temperature allow for up to 60 watts to be dissipated on the CPU, if you throttle after the 45W is reached you are leaving 15W (and thus higher clocks) on the table.

This is also the reason why MacBook Pro was known to be fast in the CPU department. By implementing their own solution, Apple could extract a bit more performance at the peak. For post Skylake CPUs it did't work anymore, since their TDP was essentially off the charts anyway.

 

[winterny]

@leman is correct.