CANT PEN INTEL POWER GADGET PRO
CANT PEN INTEL POWER GADGET CODE
CANT PEN INTEL POWER GADGET MAC

So running twice the amount of code on E cores alone takes the same amount of energy, according to Activity Monitor. Running two threads has a total energy of 1940, at 970 per thread. The single thread has a total energy of 1931 units.s, at 1931 per thread.

2 threads on 2 E cores took 10.0 s, at 194% CPU and an energy value of 194.

1 thread on 2 E cores took 19.5 s, at 98% CPU and an energy value of 99.

Then repeat that with two threads imposing the same computational load. With an app like AsmAttic, run a single thread at minimum QoS so it runs only on the E cores.

CANT PEN INTEL POWER GADGET PRO

If you’ve got an M1 Pro or Max available, it’s not hard to see where this is going wrong. It’s even worse that Activity Monitor’s errors are discouraging developers from making better use of the cores in M1 chips. Given that it’s now nearly 18 months since Apple started shipping its first M1 Macs, you might think that a little surprising. This result occurs because Activity Monitor, currently version 10.14 in macOS 12.3.1, doesn’t know the difference between processors with identical cores running at fixed frequency, and Apple’s M1 chips, with two different types of core and variable frequencies for each cluster of cores. And if you believe that, you drop the idea of offering the user control over QoS, and run all that app’s code at high QoS after all. The clear conclusion is that running these eight threads on the E cores was considerably less efficient than running them on the P cores. On 8 E cores, an energy value of 194 was sustained for 40.4 s, giving a total of 7838 units.s, or 980 per thread.On 8 P cores, an energy value of 800 was sustained for 6.6 s, giving a total of 5280 units.s, or 660 per thread.What do you get in return? According to Activity Monitor’s Energy pane: So there’s a big performance hit from constraining that code to the E cores. The results from the two different QoS settings are:

CANT PEN INTEL POWER GADGET MAC

To understand what they’re seeing, I used my app AsmAttic to run tests with different numbers of threads at the two extremes of QoS: 9 or ‘background’, which constrains the code to E cores, and the highest of 33, which runs the code preferentially on the P cores until they’re fully loaded, then uses available E cores as well.įor this introductory example, I use 8 threads of floating point maths on an M1 Max (in a Mac Studio), and a tight loop run 1 billion times in each thread. This article explains why, and its message is not to trust Activity Monitor over CPU or Energy figures. In many cases, these appear to demonstrate that running code exclusively on E cores uses more energy, not less. As more developers are looking at giving the user control over which cores do the heavy lifting in their apps, when running on M1 Macs, they’re puzzling over contradictory figures given by Activity Monitor.

Cant pen intel power gadget