Turning away from the discrete GPU to the integrated graphics in Geekbench, the results are a lot closer. The Dell still wins, though, with IGP performance about 10 percent faster.
Blender is an open-source popular rendering app used in many indie movies. It's maintained on both MacOS and Windows, but performance, unlike in Cinebench, can be uneven across OS versions. For example, rendering is generally faster on Windows 7 than Windows 10, and I've found previous builds of Blender ran faster on MacOS.
The current version seems to be pretty even-handed. Using the free BMW benchmark for Blender, the XPS 15 had a nice 10-percent or so performance advantage over the MacBook Pro 15.
Blender also supports using the graphics chip to render 3D. For this test, I tasked both laptops with a GPU render on the GPU version of the Peter Pan BMW test. The XPS 15's GTX 1050 finish quite swiftly. The MacBook Pro 15 was much, much slower—so much slower, in fact, that it's pretty apparent the GPU rendering on Blender just doesn't work right. Again, I'd blame Blender first rather than the Apple, but if you have to do GPU renders: Skip the MacBook Pro 15 for the XPS 15.
As PC gaming grows in popularity, so does the importance of graphics performance in laptops. To test these two, I chose Tomb Raider and Middle-earth: Shadows of Mordor. Both were installed from Steam onto the laptops' respective native operating systems.
The results were downright ugly. First up is Tomb Raider running at 16x10 resolution on High. The Mac pushes about 47 fps which is OK until you realize the XPS 15 is buzzing along at 137 fps. My guess is 19x10 on Ultimate is well within reach for the XPS 15. You're basically gassed out at 16x10 with the MacBook Pro 15, so getting to a higher resolution would mean compromising on even more visual quality settings.
The situation is worse for Shadows of Mordor. Mind you, I wasn't able to run the game at the exact same resolutions, so for High I opted for 1536x864 on the MacBook Pro 15 and 1680x1050 on high on XPS 15. That's about 1.32 million pixels being rendered on the MacBook Pro 15 vs. 1.76 million pixels on the XPS 15, so the XPS 15 is actually doing about 30 percent more work than the MacBook Pro 15.
One other caveat you should note for both games being tested here: This isn't a pure test of the GPU or CPU in either system, but also a test of the underlying OS, graphics API, driver, and the game itself.
Rather than see this as a dig on the Radeon Pro chip, you can see this as a dig at Mac-based gaming in general. Basically folks, this is what you get.
Yes, I know you can install Windows 10 (not free) on the MacBook Pro and play games that way, but this would mean Windows 10 is superior to MacOS, and I don't think anyone ever wants to admit that. The Shadows of Mordor performance is simply atrocious. I can run Shadows of Mordor on the XPS 15 at 19x10 on the Ultra setting and still see a very playable 48 fps, while the MacBook Pro 15 has to step down resolution (and game settings) to be even approachable to playing. Just ugly.
Winner: XPS 15
Laptops are laptops because sometimes you do, indeed, run them on battery. While you've heard on the Internets that the MacBook Pro 15 has terrible battery life, the truth is it doesn't.
For this test, we played a 4K resolution video at about 255 nits in brightness with the Wi-Fi off until the laptops died. The MacBook Pro 15 wins this one hands-down, playing the video for about 9 hours before tapping out. The XPS 15 taps out at about 5.5 hours.
What's more impressive about those results is their relation to battery capacity. The MacBook Pro 15 packs a fairly small 74-watt-hour battery, while Dell has upped the capacity on the XPS 15 to 97 watt-hours for this year's refresh.
There are mitigating factors, of course. For the most part, the GPUs are likely not part of the battery life equation, as the video playback of the file is run on the CPU's graphics chip. Skylake and Kaby Lake probably consume about the same amount of energy doing this simple task.
The Dell XPS 15, does, however have a far denser 3840x2160-pixel screen, versus the MacBook Pro 15's 2880x1600. In PPI, that's basically about 226 PPI on the Mac vs 293 PPI on the PC. Lighting up more pixels costs you more power. The XPS 15's 10-point touchscreen also absorbs some power. Other incidental system power draws, such as the SSD's, may also come into play here.
Regardless, you can't explain it away when you have 10 percent left on your battery and you're only half way through that project. The MacBook Pro 15 is the easy winner here.
Winner: MacBook Pro 15
The MacBook Pro has a few things going for it: It's lighter, and it has long battery life under light loads. But in just every other metric used to judge a performance laptop (which quad-core laptops fall into), it's the Dell XPS 15 that comes up on top. When you factor in the massive cost savings (the top-end XPS 15 cost almost the same as the base MacBook Pro 15), this entire exercise shows this rivalry continues to be entirely one-sided.