Gaming on a Ryzen 9 4900HS Laptop with a Discrete GPU

Following up to our review  of AMD's new Ryzen Mobile 4000 laptop CPU and having covered productivity performance and touched on gaming using the integrated GPU in our initial review, now it’s time to tackle the other main use case for these processors, and that’s gaming with a discrete GPU.

Image source - TECHSPOT / Image by Tim Schiesser 

This is particularly important for Ryzen H-series parts as gaming laptops stick to these 45W-ish processors almost exclusively. We know that AMD has a very compelling part on their hands, with much better efficiency plus better multi-thread and single-thread performance for the most part, but gaming can often be a story of its own. Latencies, boost performance, frequency, can all play a part, so today we’ll be doing our best to explore how the Ryzen 9 4900HS fares in games.

The biggest challenge for this test was getting an apples-to-apples platform to compare the 4900HS against Intel processors. Our Ryzen 4000 test bed, the Asus zephyrus G14, comes with a GeForce RTX 2060 Max-Q with a 65W power limit, which we haven't seen in any other laptops.

We wanted this to be a completely fair battle with the same GPU, but at least we managed the next best thing. We've put the 4900HS up against the closest platform we could find, which is a Core i7-9750H laptop with an RTX 2060 at 80W. Unfortunately we were unable to source an 8-core Intel laptop with the RTX 2060, but that will prove easier with the upcoming 10th-gen Intel line-up that features the 8-core corei7-10875H

As part of today's tests, we'll be looking at some 1080p results -- CPU and GPU limited scenarios to see what the situation is there -- and then dive into some heavily CPU limited gaming at 720p. Along the way we’ve tried to do our best to look specifically at the impact of the CPU in games. 720p creates a CPU bottleneck with the RTX 2060 in some titles, and that allows us to isolate the impact of the CPU.

We know most people will be gaming at 1080p, so let’s look at that first...

Benchmarks

We decided to start with a few very GPU-limited environments, so we can get a baseline look at the differences between the RTX 2060 Max-Q and the non Max-Q models. Control at 1080p High settings is very GPU demanding and here we see the RTX 2060 performs 9% better than the 2060 Max-Q on average, and 7% better in 1% lows

 brief intermission

At this point, we have a pretty good picture of how these systems perform. In GPU limited scenarios, the Intel combination pulls ahead due to its faster GPU. However in many games, while average performance is better on the Intel side, AMD produces better 1% low performance, sometimes this is only marginal in the 2-3% better range, other times the results are over 10% in favor of AMD.

There were also two games: Watch Dogs 2 and Resident Evil 2, where the title was CPU limited enough as a whole at 1080p to deliver better performance on the Ryzen 9 4900HS system despite its weaker GPU option.

While these results are a realistic reflection of how these combinations perform, we’re not entirely CPU limited in many of these benchmarks. On laptops running at 1080p is a bit of a borderline configuration between a GPU or CPU limit depending on the game and hardware you have. And that’s why we wanted to test 720p performance. Let’s really CPU limit these systems and see which configuration stacks up better when we run into CPU bottlenecks.

Testing at 720p to simulate CPU limited scenarios

In Grand Theft Auto V, the benchmark pass is fully CPU limited at 720p, so we run into a situation where now, the Ryzen 9 4900HS configuration is 5% faster on average and 10% faster in 1% lows versus the Core i7-9750H. This is a flip on what we had previously at 1080p, where the Intel configuration was faster on average.

Takeaways

That’s a decent chunk of benchmarks looking at a range of test conditions with… well, the best test conditions we could manage given the limitations of testing with laptops that vary so much in hardware. There are some interesting results here to break down.                                                                                                      When looking at GPU limited titles, we didn’t learn much. Understandably, the Intel system with its faster RTX 2060 GPU performed better. This will always be the case when comparing CPUs for gaming: if the title you’re playing is not limited by the CPU or bottlenecked, then whatever GPU you have is much more important and becomes the limiting factor. So you’re not going to get better or similar performance out of a Ryzen laptop with a weaker GPU when GPU limited. Makes sense.

However in a lot of the games we benchmarked at 1080p with ultra settings, the Ryzen 9 4900HS did produce better 1% low performance than the Core i7-9750H in our Intel system. The difference was only marginal in many cases, 2-3%, but given our Ryzen laptop was paired with a weaker GPU, this suggests the Ryzen 9 4900HS is more powerful in areas of our benchmark passes that are more CPU demanding.

And this follows through to when we’re seeing very CPU limited gaming scenarios. A couple of times this was the case at 1080p in games like Watch Dogs 2, Resident Evil 2 and CS:Go were all quite CPU limited at 1080p, and in each of these instances the Ryzen 9 4900HS performed better.

Then at 720p we saw a significant swing in favor of AMD. With many of the titles we looked at becoming totally CPU limited at this resolution, the Ryzen 9 4900HS delivered anywhere from 5 to upwards of 25 percent more performance on average in these games. When combined with the 1% low performance we saw at 1080p, this suggests that the Ryzen 9 4900HS is the faster gaming CPU when the GPU is taken out of the equation.

These results aren’t overly surprising when we look back to our productivity benchmarks. The Ryzen 9 4900HS ended up 5 to 15 percent faster in most single or lightly threaded applications, and we know most games these days are still more lightly threaded than heavily multi-threaded. However, in some of the best case scenarios, like Resident Evil 2’s 1% low performance at 720p which was 39% higher on Ryzen, we are more in the realm of those multi-thread results.

There are plenty of caveats to this testing though. We’ve stressed the differences in GPU countless times already, but we think the other obvious one is that we're comparing the Ryzen 9 4900HS to the Core i7-9750H. While these CPUs are found in similarly priced laptops, Intel does have 8-core Core i9 processors in their 9th-gen family and an upcoming 8-core Core i7 option in their 10th generation. In particular, a part like the Core i7-10875H does produce higher single core frequencies than the 9750H, so these results might change when we can start talking about 10th-gen benchmarks and performance.

At the end of the day, what we've learned is that Ryzen 4000 is very capable for mobile gaming. If we had two laptops that were otherwise identical aside from the CPU, the Ryzen 9 4900HS should deliver either an equal or better gaming experience than the Intel Core i7-9750H depending on how GPU or CPU limited we are. The more CPU limited, the more Ryzen benefits relative to the Intel option.

We think its reasonably impressive for AMD to jump into this market segment and provide such a competitive alternative against Intel’s most popular 9th-gen gaming laptop CPU. Intel seems to be laser-focused on frequency, while AMD is catering to that front while delivering killer productivity performance as well.

It’s also impressive to see this sort of performance at 35W, not 45W like with the 9750H we tested. That extra 10W of thermal headroom is crucial while gaming. It could translate into an extra 10W of power allocation available for the GPU in a given design, which in many situations could deliver 5 to 10 percent higher frame rates.

All of this sets us up for an interesting battle between Ryzen 4000 APUs and Intel’s 10th-gen H-series including that new 8-core. Stay tuned as new systems keep coming in for more testing in the weeks ahead.