Let’s find out what the performance differences are between the new Nvidia RTX 2070 Super Max-Q and the older non-Super version.Starting with the specs, the key difference is that the newer 2070 Super version has 11% more CUDA cores, but slightly lower base and boost clock speeds as a result.
Both have an 80 to 115 watt power limit range, basically anything not at 115 watts or max performance is considered Max-Q, and both laptops I’m testing with here run the GPU at 90 watts.
The laptops I’m testing with are the Aero 17 from Gigabyte. The one from last gen has the Intel i9-9980HK CPU and Nvidia RTX 2070 Max-Q graphics, while the newer one has the Intel i7-10875H and new Nvidia RTX 2070 Super Max-Q graphics. Yes, there’s a CPU difference, but it’s quite subtle given they have the same core count, same cache and boost similarly in this chassis, this is the closest I can do with what I’ve had available.
The 10th gen platform also offers higher memory speeds which may affect some results too, but again this is a legitimate difference that exists when comparing last gen with current gen, Super graphics will only be paired with 10th gen processors.
Let’s start out with games, then look at other 3D applications afterwards. Rainbow Six Siege was tested using the game’s built in benchmark tool. I’ve got the Super results shown by the red bars, and non-Super results shown by the purple bars. In this test the newer Super laptop was 10.6% faster with ultra settings when compared against the older non-Super laptop. Battlefield 5 was tested in campaign mode running through the same section of the game on both laptops. In terms of average FPS, the 2070 Super was ahead in all cases, with a 10% lead at ultra settings, however it was behind in terms of 1% lows. I’ve found 1% low performance inconsistent in this game, but this could also be due to our CPU difference. Shadow of the Tomb Raider was tested using the games benchmark tool, and this game saw one of the largest differences out of all titles tested, with the newer Super GPU almost 19% faster in average frame rate at highest settings. CS:GO was tested with the Ulletical FPS benchmark, and with everything maxed out the Super laptop was 27% faster than the older one, making this the biggest improvement out of all 13 games that we’re looking at here. Dota 2 was tested playing in the middle lane, and there was a much smaller difference here. 1% lows were either the same or went one way or the other slightly, while the averages were quite close too, with ultra settings just 3% faster from the Super laptop. The Division 2 was tested with the game’s built in benchmark, and there was less and less of a difference stepping up to the higher setting levels. At ultra settings, the Super laptop was just 3% faster than the older non-super version, however there was a larger 13% boost with the newer laptop at low settings, granted that’s probably more likely to be the CPU and memory differences. Overwatch was tested in the practice range as this allows me to perform the exact same test run, however actual gameplay does perform a bit worse. At max settings the Super laptop was 8% faster than the older one, though the 1% low differed much more at low settings. Assassin’s Creed Odyssey was tested using the games benchmark tool, and the Super laptop was 9% faster in terms of average FPS at the highest setting level here with the same 1% low result from either. Metro Exodus was also tested using the games benchmark, and was another win in all cases for the Super laptop, which was 18.5% faster with the extreme setting preset. The Witcher 3 was tested in the same section of the game, however the older laptop was actually slightly ahead at low to medium settings in average FPS, granted it was extremely close and negligible at the end of the day. At ultra settings where the GPU is presumably utilized more, the Super laptop was almost 4% faster. Far Cry New Dawn was tested with the games benchmark, and the Super laptop was 9% faster in average FPS at ultra settings here, while the 1% low was much closer, at least at ultra, the gap seems to widen there at lower settings. Ghost Recon Wildlands was tested with the benchmark tool, I stopped testing this in favour of the newer Ghost Recon Breakpoint, however that game launched after I did my 9th gen Aero 17 testing, so I’ve revisited it with the newer Super version. At ultra settings the Super laptop was 12% faster, with nice improvements to 1% lows seen at all setting levels. Likewise, I’ve brought back strange brigade for the same reason, I needed to pad out some games as a lot of the newer games I test these days weren’t available when I tested the non-Super laptop. In this test the Super laptop was 16.5% faster in average FPS at ultra.
These are the differences in performance when looking at all 13 games tested. On average out of these titles the 2070 super max-q was 11.6% faster than the older 2070 Max-Q, at least when both have the same 90 watt power limit. While results can vary a fair bit between different titles, as you can see here, I found it interesting that this 11% average improvement basically matches the 11% increase to the CUDA core count that the Super version has over the older non-Super.
Here’s the difference we’re looking at with 3DMark the Super laptop was 12% faster in terms of the Firestrike graphics score, almost 10% faster for the Timespy graphics score, and then it reached a 14% higher score in the port royal test which uses ray tracing.
I’ve also run SPECviewperf 13 which is a benchmark for measuring graphics performance based on professional applications. In all cases the 2070 Super laptop was ahead, however the amount could vary a fair bit between specific tests.
Luxmark is an OpenCL benchmark tool which renders out different scenes with the GPU. Depending on the test, the Super graphics performed 18 to 30% better in these tests.
The V-Ray benchmark was also run on the GPU, and it renders out a scene as well. In this case the Super laptop was scoring 20% higher in this task, so looks like the GPU specific rendering tests see further increased performance compared to most of the games.
Both laptops have the same 94wh battery, and interestingly the newer Super version lasted just a few minutes extra when compared to the older model while gaming, which may be down to some of the efficiency improvements Nvidia have mentioned with these new GPUs. Unfortunately I haven’t got power draw and thermal data for GPU load as I didn’t test this with the non-Super Aero 17 when I had access to it. Again all testing here was done with 90 watt power limits, so expect worse performance with 80 watt configurations, or likewise more performance with a higher power limit. Unfortunately the power limits aren’t advertised with new laptops.
I can’t actually find the price of the 9980HK and 2070 Max-Q version of Aero 17, but when I reviewed it late last year it was around $3200 USD. Meanwhile, this new version which is outperforming it in at least GPU tasks with the new Super graphics is actually about $300 cheaper, so it seems more performance is available for less - granted I’m not too sure how the CPUs stack up, but from what I’ve seen so far the i7-10875H isn’t far behind the i9-9980HK. Obviously if you already have a laptop with RTX 2070 Max-Q graphics, then it wouldn’t be a good use of money to upgrade to 2070 Super Max-Q, but if you’re buying today and the price gap is say a fair 10%, then the newer one does offer some improvement, and outside of games this appears to be even higher with the Super GPU in 3D rendering applications. Let me know which of these GPUs you’d pick and why down in the comments, otherwise I look forward to reading all the Max-Q hate comments.