AMD Ryzen 9 4900HS Review

Mobile computing has become AMD's next big target with its new series of APU Ryzen 4000. The new processors will arrive as the U series for low-power and ultraportable laptops and in the H series for high-performance laptops . Today we have the first retail Ryzen 4000 laptop on the market to evaluate, so the focus will be on benchmarks, so we can see for the first time how Zen 2 switched to mobile batteries compared to Intel's proven line.

All Ryzen 4000 APUs are based on the Zen 2 architecture, built using 7nm technology with a monolithic chip which also integrates a Vega-based GPU. Earlier this month, we released a full breakdown of the Ryzen 4000 APUs, detailing some of its architectural features and specifications. We will ignore most of these details in this review, but you should definitely check this out for a glimpse of what will happen in 2020.

Just to be clear, Ryzen 4000 parts do not use AMD's Zen 3 architecture. The name is a bit confusing here - Ryzen 3000 for the Zen 2 desktop chips and Ryzen 4000 for the Zen 2 APUs - but it's just a continuation of what AMD started in 2017 when they launched the first Zen APUs under the name of Ryzen 2000. APU Ryzen 4000 are therefore based on Zen 2 cores, identical to the latest Ryzen desktop processors.

It is no coincidence that the Ryzen Mobile U and H series reflect what Intel offers for different types of laptops. With the ultraportable output on hold today, it's all about the mighty H-Series and how AMD will behave for content creation, productivity and gaming.

The AMD Ryzen 4000 H series consists of three distinct references: the Ryzen 5 4600H provides 6 cores and 12 threads for the general public, the Ryzen 7 4800H bump up to 8 cores and 16 threads, alongside a variant 8-core higher in the Ryzen 9 4900H. All of these APUs come with a 45 W TDP and are complemented by a 35 W low power variant for laptops which also adheres to the rest of the AMD HS series design guidelines.

You'll also find different GPU configurations for each SKU, using Vega's refreshed design with up to 8 compute units and higher clock speeds. It's worth noting, but we expect that most H-series laptops also include a discrete GPU, like our test system. The cache, although listed at 12 MB, is actually a combined L2 and L3 figure: all parties have 8 MB of L3 cache, Ryzen 7 and 9 get 4 MB L2 and Ryzen 5 gets 3 MB.

The chip we are comparing today is the Ryzen 9 4900HS, a 35 W benchmark that provides slightly higher base and clock speeds than the Ryzen 7 4800H below 45 W, a bit of magic of binning in play there. This is not the flagship APU in the AMD range, but it will give us a really solid overview of the performance of the Ryzen 4000. We will also give you an overview of the performance of the Ryzen 7 4800H from an engineering sample, as we review the rest of the processors in the lineup as new laptops hit the market.

Our benchmark benchmark is the Asus Zephyrus G14, a neat 14-inch portable gaming laptop that incorporates Ryzen 4000 HS APUs and GeForce RTX 2060 Max-Q graphics cards. We received the most efficient model for testing, with the RTX 2060 Max-Q and the 4900HS, 16 GB of DDR4-3200 memory and a 1080p 120 Hz screen.

The focus on this review is purely the performance of the processor, so we will not go into the design and other features of the Zephyrus G14. We will say however, after a week of use, we think it is a very well built laptop, with an excellent keyboard and as we will see, very convincing performance in such a small form factor.

You will also see in this review a number of different laptop processors in graphics and sometimes GPU configurations. The data in the graphs is an average of the laptops we tested with the given hardware. Testing laptop components is naturally a bit more difficult than desktop computers, as each configuration can vary in terms of cooling and hardware, so these averages are intended to illustrate the performance of a "typical" system. Averages do not include single channel memory systems or any other situation
 that strongly strangles the components at hand, we have done our best to create apple-to-apple data whenever possible.

Landmarks
Let's start with a classic performance benchmark: Cinebench R20. In this test, we see the total domination of the red team. Not only is the 4900HS the fastest laptop processor we've tested in multithreaded workload, it's also the fastest in single-wire performance. The 4900HS is 35% faster than the 8-core Core i9-9880H when using all cores, and 7% faster in single-core. There's the Core i9-9980HK that we haven't tested which might be a bit difficult, but the 90 W figures from the 9880H suggest that it could be a difficult task. Also consider that these results come from the 4900HS at lower TDP, the 4900H at 45W should be another step forward.

There are other brutal results in this Cinebench chart. The 4900HS overwrites the Core i7-9750H, which is the most popular processor used in thin and light gaming laptops. The Zen 2 offer is more than 60% faster in the MT test. We also see that the AMD part of the previous generation is a little humiliated, the 4900HS is so much faster than the Ryzen 7 3750H in the same power envelope that it is not funny.

We also tested with the Cinebench R15 heritage which has results from a wider range of CPUs going back several years. The 4900HS remains firmly placed at the top of the graphics as expected. Anyone upgrading from a Core i7-7700HQ, for example, who was a popular H-series processor in 2017 will be treated for more than double the performance with a Ryzen 9 4900HS.

One area in which Zen 2 is receiving a particularly significant upgrade is how it handles AVX-256 instructions. It is simply much faster for wide floating point operations than before. The Ryzen 9 4900HS therefore receives a huge performance improvement in our Handbrake x265 test, which uses AVX instructions. The 4900HS is a monstrous 179% faster than the 3750H in this workload, which is simply unfathomable for a performance jump of a single generation.

AMD vs Intel side, it is also very favorable for AMD. Not quite the same margins we saw in Cinebench, but we still see a 23% performance advantage from the 4900HS over the 9880H, and a 45% advantage over the 9750H. Frankly, these are massive deltas for a laptop form factor that often receives improvements from generation to generation in single digits.

For Intel to associate the 4900HS with its 8-core offering, it must blow up its power target out of the window and instead use a 90 W TDP, which is possible on some gaming laptops with a " Turbo ”or similar. The difference in power consumption on the wall for these two systems is incredible: the G14 with the 4900HS worked comfortably at around 66 W in the long term, against 150 W for the increased power of 9880H in our HP Omen 15 test system. This shows how much more efficient AMD Zen 2 design is for these long-term workloads.

Blender tells a similar story, with huge performance gains of around ~ 35% for the 4900HS compared to the 9880H, and ~ 65% for the 4900HS compared to the 9750H. Most Blender users will likely render to the GPU instead, as it tends to be much faster, but this is another benchmark that illustrates long-term multicore performance on these laptops.

In terms of decompression, we see AMD retreat with a significant victory in 7-Zip. Ryzen processors are known to work very well in this workload. However, it lags behind the 9880H in terms of compression, up to ~ 8%, although AMD beats parts of lower number of cores like the 9750H.

The performance of Adobe Photoshop is interesting. In our Iris Blur test which is mainly limited by the processor, the 4900HS manages to beat the 9750H but remains behind the 9880H, sitting comfortably between these two processors.

This continues to be the case in the more comprehensive Puget Photoshop benchmark, which goes through a series of tests. In these two workloads, which apply various effects to very high resolution images, the 4900HS is about 10% slower than the 9880H in a heart-to-heart battle. Certainly, we see performance 40% higher than the latest generation of AMD APUs.

Now let's look at the PCMark 10 numbers. We're looking at Essential workloads
and Productivity because they are limited in CPU, while the other tests rely more on the GPU.

The Essentials workload covers things like application loading, web browsing, and video conferencing. AMD manages to match the performance of the 9880H, which may not seem so impressive, until you realize that the latest generation 3750H is down in this test. AMD is now at par performance for these daily workloads.

On the Productivity workload, we see something similar to what was happening before, with the Ryzen 9 4900HS surpassing something like the Core i7-9750H. These types of tasks were less than decent compared to the previous Ryzen, so even if the Ryzen 4000 doesn't overwrite Intel here, performance parity is a good result.

However, there are certain situations with productivity workloads where Ryzen 4000 is not as impressive. Our custom Excel benchmark offers many calculations on a large dataset, and here the 9880H again outperforms the 4900HS. The 4900HS is faster than the 9750H, so it's not a terrible result, but it seems that the large workloads of the dataset are a weakness for Ryzen.

MATLAB allows us to reconfirm this with our ODE and FFT benchmarks. Again, a lot of data is crunching on big data sets and the Intel 9880H comes away with a win.

Another workload where Ryzen works well, but not enough to beat competitor Intel 8-core is our Acrobat PDF export test, which is fully single-threaded. The 9880H is slightly faster in Acrobat, although we still see a significant increase in performance for the 4900HS compared to the 3750H with a healthy 20% increase in performance for a single thread.

A final workload before examining GPU acceleration is the AES-256 performance provided by the Sandra benchmark from SiSoft. In the multithreaded test, we see 15% better AES performance compared to the 4900HS compared to the 9880H, making it the fastest processor we've reviewed for cryptographic workloads. This gives Ryzen a clear advantage in two low-level tasks heavily used in decompression and cryptography.

Now let's move on to our Premiere tests which are mostly GPU accelerated, starting with our 1-pass encoding that takes advantage of Intel QuickSync technology. Premiere doesn't support hardware accelerated encoding on AMD processors at this point, so for those who like fast export with slightly reduced image quality in the end, Intel is still the way to go. In particular, the 4900HS associated with Nvidia's GeForce RTX 2060 Max-Q is around 20% slower than a Core i7-9750H with an RTX 2060 simply because hardware acceleration is not available.

On the other hand, our Core i9-9880H system does not support QuickSync acceleration because the iGPU is fully disabled to support G-Sync via the Nvidia GPU. We therefore have a nice comparison of software coding between the 9880H and the 4900HS, where the 4900HS strongly withdraws by up to 26%.

Then we come to our 2 pass coding which produces superior image quality and does not support hardware acceleration. In this scenario, performance comes back to what we've seen in most long-term workloads: the 4900HS is 30% faster than the 9880H and 38% faster than the 9750H, despite the fact that our laptop 9880H incorporates a much faster GPU. This test is not limited by the GPU, but even with a discrete GPU of moderate capacity, it seems that AMD's Zen 2 APU is much faster for Premiere encoding.

But wait, there are more Premiere tests. Here we have a single instance of Warp Stabilizer, which is an extremely demanding effect that runs on a single thread per instance. The 4900HS manages to stabilize the footage 14% faster than the 9880H and 22% faster than the 9750H, which shows the power of the high performance single wire that we saw for the first time with Cinebench.

And finally, we have the Puget benchmarks. The Ryzen 9 4900HS is the best laptop processor for live playback in Premiere, with 12% better performance than the 9880H, which means we are seeing better editing performance. Next, for their export benchmark, we see similar numbers to the ones we just talked about, with QuickSync acceleration helping some of that workload.

Game benchmarks
Although we do not think that much games
 will actually use the GPU built into these H-series processors since the vast majority of H-series laptops also include discrete graphics, it's worth taking a look at the performance of the iGPU in a small selection of low-intensity games, just to see what AMD improvements have been made to the Vega GPU.

The Ryzen 9 4900HS sports the fastest iGPU configuration available in the series with 8 computing units clocked up to 1750 MHz. And that gives impressive results. In Grand Theft Auto V, the 4900HS provided 36% more performance than the Ryzen 7 3750H, which included 10 Vega computing units at 1400 MHz. On paper, the two processors have similar raw GPU performance, but with all the benefits that AMD has talked about, such as increasing memory bandwidth, this new Zen 2 APU is capable of moving forward.

In Civilization VI, the 4900HS again offers 37% more performance than the 3750H in the same 35 W power envelope. And by switching to CS: Go, this margin remains the same at around 37%. The 4900HS running CS Go exclusively on the integrated graphics card was able to achieve over 100 FPS on average using low settings, which is very nice.

And finally, we have our most performance-intensive iGPU test in Gears 5 running at medium parameters. The 4900HS is about 31% faster than the 3750H in this workload, making it faster than a low-end discrete GPU like Nvidia's MX250.

It is not common to see an H-series laptop paired with an MX-class GPU. Usually, OEMs opt for more powerful configurations like a GTX 1650 or higher, but with the Ryzen 4000, there really is no need to bother with an MX250 or similar discrete GPU in an H-series design .

Durable clocks and other issues
At this point in the review, we will normally guide you through comparison summaries between the Ryzen 9 4900HS and various other processors. But we still think there are a few performance questions left on the table. The first is how fast the processor actually runs in practice, and how the boost behaves. And the other is, why do we see a drop in performance in some heavy data workloads like Matlab, Excel and Photoshop. So let's start with this second part first.

We have two work theories to explain why we see heavy data workloads working this way. The first is simple, based on the specifications of these processors, especially the cache sizes. The Core i9-9880H has a decent 16 MB of L3 cache, which is what Intel offers on the desktop with parts like the i9-9900K. However, the Ryzen 9 4900HS only has 8 MB of L3 cache, half of what Intel offers, and well below the 32 MB of L3 that AMD integrates into their 8-core Zen 2 desktop processors like the Ryzen 7 3700X.

Having less cache means that less data can be stored in super fast memory and accessible in an instant. When you have a high number of cores, high performance processor cores but not enough cache, this can become a bottleneck in some cases. And although this amount of cache has doubled on Ryzen 4000 compared to previous mobile series, the same goes for the number of cores. It's probably not the whole story, but part of it.

The other is the memory system. Yes, AMD offers higher memory bandwidth than Intel with support for DDR4-3200 speeds, Intel only offers DDR4-2666 with the 9th generation. In a benchmark like Sandra, AMD provides approximately 35% more memory bandwidth. However, Ryzen 4000 seems to have lower memory latency. This, like the size of the cache, can become a performance constraint. The Core i9-9880H has a memory latency of around 30 ns for data sets larger than 32 MB, while the Ryzen 9 4900HS has a memory latency of 46 ns. This is a substantial victory for Intel.

It's hard to say for sure if these factors are part of the cause, or the whole cause, but looking at low-level benchmarks are the two things that stood out the most for us. When we come to compare more Ryzen 4000 APUs, we will have a clearer idea of ​​where these bottlenecks are.

As for clock speeds and boost behavior, let's take a look. After a cold start in our Handbrake AVX workload, the Ryzen 9 4900HS has steadily increased to around 65 W of power for some
seconds, reaching 4 GHz boost clocks every core, before falling to 54 W for a longer period, with clocks of around 3.7 GHz. Finally, the CPU settles down at 35 W to provide 3.2 GHz whole-hearted, just above the base clock of this 3.0 GHz processor. The boost period can vary depending on the heat of the system, but from a cold start, we generally observed at least 2 minutes of boost of ~ 53W, which is generous. Temperatures were actually very well managed on the Zephyrus G14, with its air cooler providing a tick of over 70 ° C sustained, but of course that will vary between laptops.

Performance summary
Ryzen 9 4900HS vs Core i9-9880H
The big one here is the Ryzen 9 4900HS compared to the Core i9-9880H. These are similar class processors, 8 cores versus 8 cores and should give us a pretty solid indication of what to expect from the 10th generation, if we get minor improvements in clock speed from Intel, like we say it.

In most long-term and multi-threaded workloads, we see 35% better performance from the Zen 2 chip. We also get impressive numbers for video encoding. Single-threaded workloads are also generally faster, or in the worst case like light productivity, equivalent to that of Intel. However, performance lags behind heavy data repositories like Excel, Photoshop and Matlab.

The amazing thing about most of these results is that AMD is able to achieve, in many workloads, at least 25% better multi-thread performance while providing lower sustained power consumption, at 35 W versus 45 W. So what happens when you limit the Intel 9880H to a PL1 of only 35 W using Intel XTU software?

Well, the margins are increasing even more. Although we let these processors do what they normally do in the boost phase, some benchmarks don't change too much, long-term workloads are now strongly in favor of AMD. Performance above 50% is possible in certain situations, and a 40% improvement in the handbrake is impressive. This means that an OEM designing a thin and light system with limited cooling will generally get the best experience by far with Ryzen.

Surprisingly, even when we eliminate the 45W power limit from Intel and increase it to 90W forever, Intel still can't beat the 35W AMD processor in long-term multi-threaded workloads. In an astonishing victory for efficiency, the 4900HS is either equal to this 90W processor configuration, or up to 15% faster. With these two systems, we have observed a difference in power consumption of 80 W compared to the wall, so if Intel wants to match AMD with 14 nm processors by simply increasing the power limits, laptops will need much larger coolers to deal with.

Ryzen 9 4900HS vs Core i7-9750H
Comparing the 4900HS to 9750H, there is only one instance where the 9750H system can be faster, and that is with QuickSync Premiere accelerated encoding. In all other references, including those where previously the 9880H was faster, the 4900HS is now faster. This bodes well for all future battles in the Ryzen 7 / Core i7 range where AMD has a basic advantage by exclusively offering 8-core parts.

Ryzen 9 4900HS vs. Ryzen 7 3750H
Now let's see how far AMD has come with Zen 2 compared to Zen + in a 35 W mobile design. The difference in some workloads is staggering. AMD not only produces a 30% improvement in performance in single-threaded workloads, but is capable of delivering 2.5 times the performance in long-term multi-threaded tasks like Blender, Handbrake and Cinebench - with the same absorbed power.

Ryzen 9 4900HS vs. Ryzen 7 4800H
We promised a preview of the Ryzen 7 4800H's performance on an engineering laptop sample to which we had brief access. Based on a small subset of references, the performance seems equal, if not slightly better, to the 4900HS, which bodes very well for more mid-range systems which will use the Ryzen 7 part. Yes, the 4900HS is more efficient , but the 4800H should still offer 8-core performance.

What we have learned...
Ryzen 4000 provides the best performance for high productivity workloads you can get in a mobile form factor. We can say that without even testing the RyZen 9 4900H more powerful, because the 35W variant of the Ryzen 9 4900HS already leaves the 8-core Intel competitors in the dust.

The Ryzen 9 4900HS is a great choice for heavy work: we're talking about video encoding or transcoding, 3D rendering, file compression, that sort of thing. It is much faster and does so at a lower power consumption, which allows better performance in a smaller form factor. Even in single-threaded workloads, AMD usually comes out on top, which means that the Premiere editing experience is better on the Ryzen 9 4900HS, for example.

From a pure generation to generation upgrade point of view, AMD has done an incredible job. The Ryzen 7 3750H was hard to recommend and was easily beaten by mid-range parts from Intel. All of that has changed with the new APUs.

Technically, Intel parts can still compete in performance if you remove the long-term power limit of 45 W. We saw it with the Core i9-9880H and we expect it to be. even if we had tested the flagship Core i9-9980HK. But to get that kind of performance, you need a bigger laptop with a more robust cooler, and it'll probably run hotter too. The Ryzen 9 4900HS can provide all of this performance in small laptops - like this Zephyrus G14 - and at the same time run cooler.

However, AMD's Ryzen 4000 processors are not the complete package, there are some limitations. The first concerns heavy data workloads, such as computing massive numbers of spreadsheets in Excel, working with huge photos in Photoshop, or running Matlab scripts. Intel won in these tests. Lighter loads will do the trick on AMD, but if you're a big computer, the equivalent processors at the heart of Intel might be the way to go.

The other concerns workloads that use proprietary Intel technologies like QuickSync, Premiere encoding being one. If it's something you use all the time, Intel is the way to go. It's worth mentioning that all of Premiere's other tasks besides QuickSync encoding are faster on Ryzen 4000, including editing and applying intensive effects like the Warp Stabilizer.

Apart from these two instances, you will have a better and faster productivity experience with these 8-core Ryzen 4000 processors. This is unexplored territory for AMD on a form factor for mobile PC.

It also seems that AMD will offer this type of performance at very competitive prices. The Zephyrus G14 with the Ryzen 9 4900HS and RTX 2060 Max-Q is expected to cost around $ 1,500, which is firmly the territory of the six-core Core i7-9750H laptops. The Core i9-9880H laptop we bought for testing in this review cost $ 2,400 and was one of the cheapest we could find. Granted, it has a much faster GPU for gaming, but at the moment, you won't find 8-core laptops in the $ 1,500 price range. While most of our data is focused on Ryzen 9 vs Core i9, it isn't really competing with each other, at least for now, as it may change with Intel's 10th generation.

Some additional notes to conclude this review…

With previous Ryzen laptops, we did not have the most stable experience, encountering many software bugs and crashes, which we have discussed a few times before. Completely different experience with the Ryzen 9 4900HS in the Asus Zephyrus G14. Even with a beta version of the GPU driver on this laptop, we did not experience any stability issues, crashes, or applications that refused to work properly.

You may also be wondering about gaming performance with a discrete GPU. The Zephyrus G14 includes the RTX 2060 Max-Q after all. This is something that we will explore in a future review, as we are still working to get the best comparison between apples and apples. Hoping to have it soon.

Battery life will also be on people's minds, but that is not something we intend to test as we did not have enough data for various laptops to compare it. fair.

Finally, we will end this one with some comments on the 10th generation of Intel. New laptop processors are fast approaching, probably within a month. Rumors Suggest Intel Sticks to 8-Core H-Series Coins and Always
using 14nm technology. We just can't see how they can compete in terms of performance or efficiency. If all we get is a little clock retarder with the same 14nm efficiency as the last generations, Ryzen will easily win this one. We will see this battle unfold soon, but based on what we have seen so far, we have no great hopes for the 10th generation.