Reaching for Turbo: Aligning Perception with AMD’s Frequency Metrics
by Dr. Ian Cutress on September 17, 2019 10:00 AM ESTA Short Detour on Mobile CPUs
For our readers that focus purely on the desktop space, I want to dive a bit into what happens with mobile SoCs and how turbo comes into effect there.
Most Arm based SoCs use a mechanism called EAS (Energy-Aware Scheduling) to manage how it implements both turbo but also which cores are active within a mobile CPU. A mobile CPU has one other aspect to deal with: not all cores are the same. A mobile CPU has both low power/low performance cores, and high power/high performance cores. Ideally the cores should have a crossover point where it makes sense to move the workload onto the big cores and spend more power to get them done faster. A workload in this instance will often start on the smaller low performance cores until it hits a utilization threshold and then be moved onto a large core, should one be available.
For example, here's Samsung's Exynos 9820, which has three types of cores: A55, A75, and M4. Each core is configured to a different performance/power window, with some overlap.
Peak Turbo on these CPUs is defined in the same way as Intel does on its desktop processors, but without the Turbo tables. Both the small CPUs and the big CPUs will have defined idle and maximum frequencies, but they will conform to a chip-to-chip defined voltage/frequency curve with points along that curve. When the utilization of a big core is high, the system will react and offer it the highest voltage/frequency up that curve as is possible. This means that the strongest workloads get the strongest frequency.
However, in Energy Aware Scheduling, because the devices that these chips go into are small and often have thermal limitations, the power can be limited by battery or thermals. There is no point for the chip to stay at maximum frequency only to burn in the hand. So the system will apply an Energy Aware algorithm, combined with the thermal probes inside the device, to ensure that the turbo and workload tend towards a peak skin temperature of the device (assuming a consistent, heavy workload). This power is balanced across the CPU, the GPU, and any additional accelerators within the system, and the proportion of that balance can be configured by the device manufacturer to respond to what proportion of CPU/GPU/NPU instructions are being fed to the chip.
As a result, when we see a mobile processor that advertises ‘2.96 GHz’, it will likely hit that frequency but the design of the device (and the binning of the chip) will determine how long before thermal limits kick in.
Facebook
Twitter
RSS
144 Comments
View All Comments
MDD1963 - Tuesday, September 17, 2019 - link
AMD's four visits totaling little 1/16th of a second each of a single core at advertised boost clocks within a 45 - 100 secnd period are most impressive! :)realHolt - Tuesday, September 17, 2019 - link
Seen this: https://www.youtube.com/watch?v=3LesYlfhv3otwtech - Tuesday, September 17, 2019 - link
If I buy a processor that's advertised as being able to reach a certain clockspeed, I expect the individual example of that product to be able to reach that speed - not that there are some other units marketed under the same name that may be able to. Better to under-advertise than over-advertise.Korguz - Wednesday, September 18, 2019 - link
twtech and how do you feel about a cpu being advertised as using X amount of power, but using much more then that ?eva02langley - Wednesday, September 18, 2019 - link
It is not ADVERTISED as a 4.4GHz CPU either... you just made the assumption that it was!Techie2 - Tuesday, September 17, 2019 - link
IMNHO it's much nonsense over nothing. No problem with CPU turbo mode IME. Yes there was a 25 MHz drop with the first chipset update to fix overly sensitive load sensing that kept some CPUs at max vcore but other than that no issues for me and no malice or fraud by AMD.FYI - AMD settled the bogus CPU core count lawsuit as it was cheaper to settle a frivolous lawsuit than fight it in U.S. courts where Jackpot Justice prevails by criminals. Pour a cup of coffee on your crotch in a moving vehicle and become an instant millionaire. It's the American way...
zodiacfml - Tuesday, September 17, 2019 - link
Never bothered to read on this issue, surprising Andntech did. I guess its quiet right now in the tech world. This topic is of little value considering that even Intel's turbo boost speeds are not so easy to predict unless you have an application locked to a certain number of threads and CPU utilization.PProchnow - Tuesday, September 17, 2019 - link
Now like if you think Intel I want 5.5Ghz locked in low on 12 cores!!!!I got a 4.4Ghz locked SANDRA Graphics Handling Benchie that convinced me the 3900X R A W K Z !
PProchnow - Tuesday, September 17, 2019 - link
Really, i will KEEP the X470 if this is the best the whole CHANNEL consortium can show me. No huge upside fo buy into.
BTW I have been seeing the Turbo Peaks you got since i dropped the 3900X in the Friday after the Sunday release....AMD direct sale on Sunday Morning. I seem them in CPUZ HWMON max history. That sim-ly grabs the high number for you no B.S.
Peter2k - Wednesday, September 18, 2019 - link
The only important difference might be a better VRM solution, which has been improved generally speaking on x570 compared to older generations.But if you bought the right board then it's not an issue (but then you would have to know what to look for)
And even then, it really only matters once you push past 8 cores and obviously put a high load on the chip
Btw, not needing to buy a new board is a good thing