System Performance

Let’s talk performance and once again here I’ll be mostly going through how the HP Envy x360 13 performs in a summarized way because we’ve already covered the Ryzen 5 2500U in detail a fair while ago, and those results are still perfectly valid. And yes, the Envy x360 unit we were sent packs the Ryzen 5 2500U.

Also available is the Ryzen 7 2700U if you want slightly higher CPU clocks and a decent amount more GPU performance, while on the lower end you can get the Ryzen 3 2300U, although I’d recommend sticking with at least the Ryzen 5 to get a decent GPU. There’s also 8 GB of DDR4 as standard, and a 256GB SSD, although those parts are upgradeable if you’d like. Strangely, HP only offers a 16GB RAM upgrade with the Ryzen 5 SKU, not Ryzen 7. A 4K screen is also optional though I’d stick to 1080p for performance and battery life reasons.

The main difference between the Ryzen 5 2500U in the 13-inch Envy x360, and the one in the 15-inch Envy x360 we used for our original Ryzen 5 2500U testing, is the power configuration. The 15-inch model with more room for cooling uses a 25W configuration, whereas the 13-inch I’m reviewing today uses the standard 15W configuration. While the chip and its specs is identical, this TDP difference governs how power is split between the CPU and GPU, as well as what levels of boost clocks can be sustained and for how long, so naturally there’s going to be a performance difference when going from 25W to 15W.

The good news is that when you compare the two configurations, the performance difference is only significant in a few types of workloads. In shorter, mostly CPU-bound tasks like Cinebench, Excel, 7-Zip and PCMark, there is very little performance difference; either within the margin of error, or reduced by a single digit amount. Considering we’re talking about a 40 percent reduction in TDP, and most ‘everyday’ tasks are these shorter, CPU-bound type things, that’s a great result for the 15W configuration of Ryzen 5 Mobile.

Where the 2500U does take a hit is in longer workloads, specifically things like video encoding. Handbrake, for example, is 27 percent slower in a test that takes over 2 hours to complete. Blender is another benchmark that sees around that 27 to 28 percent hit, which again isn’t quite as large as the 40 percent cut to power.

The other area that gets impacted is GPU performance, and it’s there you’ll see reductions by around 30 to 35 percent, even in shorter loads. It’s no real surprise the higher 25W configuration is able to allocate more power to the GPU, so when you bring that down to 15W, it’s an area that is going to suffer a bit.

Discussing the 15W versus 25W configurations is just for interests sake really, as the main comparison is between the Ryzen 5 2500U and the Core i5-8250U, the latter of which is found in competing systems around this price. And here, Ryzen is quite competitive depending on the workload, particularly in tasks that utilize the GPU.

In Cinebench R15, the Ryzen 5 2500U takes a strong 21 percent lead and there were also handy wins in 7-Zip decompression and our Premiere non-Lumetri test. Blender was also a good 15 percent faster on the 2500U thanks to Ryzen Mobile’s far more impressive integrated graphics. Several other tests pulled up around the same level as the 8250U, including Excel, x264 encoding’s second pass, and PCMark.

The two major losses shouldn’t come as a surprise. Handbrake uses AVX instructions for encoding x265 videos, and Ryzen is well known to have weaker AVX performance compared to equivalent Intel parts, hence why it falls 28 percent behind. MATLAB, a commonly used tool for engineers, isn’t optimized for Ryzen at all and in fact uses Intel-optimized binaries which sees Ryzen suffer significantly in that test.

In GPU bound scenarios like 3DMark’s Time Spy test, the Ryzen 5 2500U holds an enormous lead on the Core i5-8250U, to the tune of more than 50 percent, and this carries across into GPU-bound games and compute-heavy applications. There is no doubt whatsoever that the Ryzen 5 2500U is far superior to the Core i5-8250U in graphics, and in some cases like Fortnite it’s actually possible to play the game comfortably on the 2500U where the 8250U is simply not fast enough.

And that’s really the tale of the two CPU options: the Ryzen 5 2500U offers equivalent CPU performance to the Core i5-8250U in most applications, except the 2500U offers significantly better graphics performance. If you want to do anything GPU related, including playing basic-level games, and you’re tossing up between the Envy x360 with Ryzen Mobile, and something with Intel inside, Ryzen Mobile is a much better option.

Both AMD and Intel have higher tier CPUs on offer in a 15W TDP configuration: on the Intel side there’s the Core i7-8550U, and on the AMD side we have the Ryzen 7 2700U, so it doesn’t seem fair to compare the Ryzen 5 2500U in my review unit to, say, the Core i7-8550U. But if you do want more performance, there is an Envy x360 13 with the Ryzen 7 2700U inside, so you’ll be able to enjoy the faster CPU and GPU that APU provides. And again, the 2700U is extremely competitive with the 8550U, particularly in graphics where Ryzen Mobile absolutely dominates.

There is one other laptop configuration we want to mention, and that’s the Core i7-8550U paired up with Nvidia’s GeForce MX150 1D12, a combo we’ve started to see in 13-inch notebooks recently to provide a decent GPU in Intel-powered systems to replace their mediocre integrated GPU. The MX150 1D12 is a fair bit faster than the GPU in the Ryzen 5 2500U, but the 2700U certainly gives it a run for its money. Also, the MX150 + 8550U combo requires more power than Ryzen Mobile, so that’s something to keep in mind for those that run workloads on battery.

We haven’t done extensive testing with the Ryzen 7 2700U or MX150 1D12, but what we'll say is don’t automatically assume the discrete MX150 is superior to the Ryzen 7 2700U’s integrated Vega GPU simply because it’s a discrete GPU. Ryzen Mobile is highly competitive with the MX150 1D12 and that’s why we think it’s such a compelling option for ultraportable systems.