Performance Analysis and Clock Speed Scaling
CPU and General Performance
In order to evaluate the overall performance of the XPS 13 2-in-1, we pitted it against similar form factor machines including the Huawei MateBook that we previously reviewed which as 2-in-1 machine featuring a Skylake Core-M processor.
In the Cinebench R15 rendering benchmark, we see that the XPS 13 2-in-1 beats the MateBook on single threaded performance, and is within 5% in the multi-threaded test.
In Handbrake, we see the Dell XPS 13 2-in-1 beating the MateBook by 13% for the CPU encoding benchmark.
PCMark 8 is a benchmarking suite that aims to emulate several different usage scenarios ranging from basic productivity to mixed workloads with light gaming and to applications for creative professionals like photo and video editing. While the "conventional" tests are running applications as you'd expect, the "accelerated" versions add OpenCL acceleration and use the available GPU devices for some operations.
In PCMark 8, we once again see the XPS 13 2-in-1 beating the MateBook, but it's still behind the higher power mobile processors in the ThinkPad X1 and MacBook Pro.
As we mentioned earlier in this review, the XPS 13 2-in-1 features a technology that Dell is calling Dynamic Power Mode. Essentially this amounts to greater control over dynamically adjusting the clock speed of the i7-7Y57 processor in order to stay within the thermal constraints of the chassis as well as improve battery life.
In order to monitor the clock speed, we used the Intel Power Gadget software to the processors over a 10-minute portion of PCMark 8.
First, we'll take a look at the Huawei MateBook.
Here we can see the MateBook scaling the CPU frequency according to what is happening in the benchmark, as expected.
Now let's take a look at the same test on the XPS 13 2-in-1.
Here we can see that the XPS 13 is more aggressive on dynamically changing the clock speed. The XPS 13 spends more time at the in the 900MHz clock speed range than the MateBook. This keeps the Dell system in its lower power state more often, while still being able to "burst" up to the 3.5 GHz range for workloads that demand it. The restrictions are on the length of time of that burst, in order to maintain sustained power consumption at a lower level to extend battery life and keep the surface temperatures low.
It's nice to look at this data, what does it ultimately mean? Well, I have to say that day to day usage of the XPS 13 2-in-1 has not presented many performance discrepancies over any other "ultrabook" machine. The "bursty" nature of many workloads is more difficult to test and show case, as the result is end-user experience rather than any hard coded performance metric. This leaves us with a situation where the benchmarks do not necessarily match up with that user experience. The Dell Dynamic power mode seems to have its advantages in real-world use, but the Dell claim of 10% better performance than the Skylake XPS 13 seems to be difficult to pinpoint.