Samsung recently unveiled Galaxy S22 series smartphones, marking a major paradigm shift in its flagship portfolio by consolidating the Galaxy S and the Galaxy Note series. Although the Galaxy S22 series did not include many hardware changes over the previous generation, performance gets a major boost with new 4nm chipsets.
Samsung’s Galaxy S22 models
Like previous years, Samsung is retaining its dual-chipset policy for its flagship Galaxy S22 series and will be selling the flagship devices with chipsets varying as per region. While major markets such as the U.S., Canada, South Korea, and Hong Kong will get Galaxy S22 smartphones powered by Qualcomm’s flagship Snapdragon 8 Gen 1 mobile platform, we can expect Samsung to task its own flagship Exynos 2200 chipset with powering the Galaxy S22 in other regions.
As per tipster Dohyun Kim, the Galaxy S22 series will feature the Snapdragon 8 Gen 1 chipset in more regions, including North and South Americas and East as well as Southeast Asia. Meanwhile, regions such as the Middle East, West Asia, and Africa will get a mix of Exynos 2200 and Snapdragon 8 Gen 1, whereas Samsung will continue to sell the Galaxy S series in Europe with Exynos chips as earlier.
Galaxy S22 Series AP by Regions
1. Europe: Exynos
2. North America: Snapdragon
3. South America: Snapdragon
4. East Asia: Snapdragon
5. South East Asia/Oceania: Snapdragon
6. West Asia: Exynos & Snapdragon
7. Middle East Asia: Exynos & Snapdragon
8. Africa: Exynos & Snapdragon— Dohyun Kim 𝕐 (@dohyun854) January 20, 2022
While we can say for sure that the U.S. will get the Snapdragon variant of the Galaxy S22, chipsets for other regions cannot be confirmed at the moment until Samsung begins shipping these devices or device reviews start pouring in from other regions.
Snapdragon 8 Gen 1 vs. Exynos 2200 on Samsung Galaxy S22
While Qualcomm’s flagship Snapdragon chipsets have traditionally overpowered Exynos chipsets, the performance gap between Samsung’s flagship Exynos chips and Qualcomm’s top-of-the-line Snapdragon chips has reduced significantly, especially since the launch of Exynos 2100 in 2021. The Exynos 2200, which was announced last month, puts up a tough fight against the Snapdragon 8 Gen 1 thanks to almost equivalent hardware. Samsung also makes major claims about its new custom GPU that has been designed in collaboration with AMD to offer more refined graphics on Samsung phones.
As expected, while Qualcomm’s Snapdragon 8 Gen 1 powers multiple flagship smartphones, including the Xiaomi 12 series and the OnePlus 10 Pro besides the North American and Korean units of the Galaxy S22, the Exynos 2200 is limited to the Galaxy S22 series. With these considerations in mind, the following sections discuss how the two rivals compare.
But before we dive in, here’s a quick comparison between the Snapdragon 8 Gen 1 and Exynos 2200.
| Chipset | Qualcomm Snapdragon 8 Gen 1 | Samsung Exynos 2200 |
| CPU |
|
|
| GPU | Adreno GPU with Snapdragon Elite Gaming features | Xclipse GPU with AMD RDNA 2 |
| ISP | Snapdragon Sight, 18-bit color depth | 16-bit color depth |
| NPU | 7th Gen AI Engine | AI Engine with dual-core NPU and DSP |
| Modem | Snapdragon X65 integrated modem with up to 10Gbps downlink speed | Integrated modem with up to 10Gbps downlink speed in EN-DC mode |
| Fabrication process | 4nm Samsung EUV | 4nm Samsung EUV |
4nm ARMv9 CPUs
The Samsung Exynos 2200 and Qualcomm Snapdragon 8 Gen 1 are quite comparable when it comes to the CPU architecture. Both Samsung and Qualcomm are employing ARM’s licensable ARMv9 CPU designs IP for their flagship chipsets.
Both also feature an octa-core design with a tri-cluster (1+3+4) arrangement of cores, using one ARM Cortex-X2 core as the primary performance core, three ARM Cortex-710 cores for tasks with middling power requirements, and four ARM Cortex-510 cores for efficiency.
There is a slight variation in terms of the core frequencies, as can be seen in the table above, and thanks to higher core frequencies, the Snapdragon 8 Gen 1 is expected to take the lead in CPU performance. However, higher core frequencies also put the chipset at the risk of overheating — which can lead to thermal throttling. On the other hand, Samsung appears to be taking the wise step of consciously limiting the clock speeds of the Exynos 2200 to avoid throttling-related issues, which have plagued the Exynos lineup of chips for several previous generations.
The Exynos 2200 and the Snapdragon 8 Gen 1 are also fabricated alike, built using Samsung’s EUV (extreme ultraviolet lithography) process on a 4 nanometer (4nm) node. Samsung Semiconductor claims the downsizing from 5nm to 4nm EUV node leads to a 16% increase in the power efficiency of the chipsets. Since the two chipsets are built in the same foundry on the same process, we can expect these figures to translate to similar advantages for both.
Notably, MediaTek’s Dimensity 9000, which also features a very similar core distribution, goes a step higher with the clock speed of the Cortex-X2 primary core. The chipset, although also fabricated on a 4nm node, is made by Taiwan’s TSMC instead of Samsung.
Samsung bets big on AMD GPU
Although Qualcomm had been long rumored to revamp the naming of its flagship Snapdragon chipset, the company did surprise us by dropping the numeral identifier from its Adreno lineup of GPUs, and the step might suggest the company has deprioritized its GPU. At the launch of the Snapdragon 8 Gen 1, Qualcomm not only rescinded a formal naming upgrade to the GPU but also spent very little time detailing the upgrades with the new Adreno GPU.
Qualcomm only emphasized the 30% graphical performance improvement brought about by the new Adreno GPU over the last generation used in the Snapdragon 888. The company did mention architectural modifications, including improvements in concurrent processing, which is likely to translate to better real-world graphical performance even if the improvements do not have a major impact on the benchmark results.
Meanwhile, Samsung is going all out with its new collaboration with AMD for the new GPU. The Exynos 2200 features the highly anticipated Xclipse GPU, which utilizes AMD’s RDNA 2 technology. Samsung claims AMD’s RDNA 2 technology in the new GPU will allow mobile game developers to add support for ray tracing on Android games.
For those unfamiliar, ray tracing is a feature that allows rays emanating from light sources to look more realistic and fall more naturally on objects in the gaming world. These realistic visuals add to the overall appeal of the gaming graphics and make the game more immersive and enjoyable.
With the advent of VR headsets that use mobile chipsets, gaming on these headsets may improve significantly as ray tracing becomes widely adopted across the mobile industry. Currently, raytracing is seen on desktops through high-end graphics cards, including Nvidia’s RTX and AMD’s RX 6XXX series.
Notably, MediaTek also announced support for ray tracing on the Dimensity 9000’s Mali GPU, but its implementation relies on software-based emulation instead of actual hardware acceleration. At the moment, we cannot verify the claims by Mediatek as there are no mobile games that support ray tracing currently.
Image Signal Processor
Qualcomm did not seem to focus on the GPU as much as Samsung, but it has brought a massive rebranding to the image signal processor (ISP) on the Snapdragon 8 Gen 1. The new ISP on the Snapdragon 8 Gen 1 is now marketed as “Snapdragon Sight,” and the company claims mighty improvements in processing in the image signal chain.
The first major advancement that Snapdragon 8 Gen 1’s new ISP brings comes in the form of improved color depth. According to Qualcomm, Snapdragon Sight is capable of processing 18-bit colors. This means the new ISP can capture 16 times more colors as compared to 14-bit processing on the Snapdragon 888’s Spectra ISP. For users, this means smartphones powered by the Snapdragon 8 Gen 1 will capture more realistic colors, wider HDR, and reduced noise in the images.
In addition, night photography is touted to improve significantly with the new ISP. To capture better details at night, smartphones capture multiple images at different exposure levels at night and superimpose them. With Snapdragon Gen 1, this image stacking can use up to 30 images for better details and brighter shots. As per Qualcomm, the ISP can process five times (5x) better images in the night mode as compared to the Snapdragon 8 Gen 1.
Snapdragon Sight uses neural networks for better auto-focus, auto-exposure, and auto-white balance. Qualcomm notes all the machine learning (ML) and artificial intelligence (A.I.) processing happens right on the ISP instead of being offloaded to another part of the chipset. Furthermore, the ISP now features a hardware engine for bokeh, meaning portrait images and videos will be much more refined. In fact, phones with Snapdragon 8 Gen 1 are capable of portrait videos at up to 4K resolution, which is comparable to the highly-appreciated Cinematic mode on the iPhone 13 series.
Lastly, Qualcomm says phones driven by the Snapdragon 8 Gen 1 will be able to capture 8K video with HDR while also being able to capture 64MP shots simultaneously.
In contrast, the list of improvements brought about by Exynos 2200’s ISP is relatively small. First of all, Samsung claims the new ISP supports image capturing at up to 200MP (even though the Galaxy S22 Ultra only features a 108MP primary camera). It can shoot videos at up 8K resolution and 30 frames per second (fps) and 4K videos at up to 120fps with HDR10+.
Samsung also claims Exynos 2200’s ISP can process signals from four cameras simultaneously. It can use A.I. to optimize parameters for better colors, auto-white balance, auto-exposure, and HDR.
Talking of color depth, Samsung says the Galaxy S22 Ultra supports image processing with 16-bit colors, although there is no such mention for the vanilla Galaxy S22 and the Galaxy S22 Plus.
Neural Processing Unit: A.I. and ML applications
Qualcomm uses the 7th Gen AI Engine on the Snapdragon 8 Gen 1 chipset. Qualcomm claims the new NPU is twice as fast as the Snapdragon 888 when it comes to machine learning-related performance. It also claims that the processing of A.I. workloads is 70% more efficient on the Snapdragon 8 Gen 1’s new NPU.
Meanwhile, the Exynos 2200 features a dual-core neural processing unit (NPU) and digital signal processor (DSP). It says that the new chipset can compute twice as fast as compared to the Exynos 2100 when it comes to A.I.-related applications. The faster NPU allows for multiple streams of computations to take place in parallel and with greater precision.
Connectivity
Qualcomm is using an integrated Snapdragon X65 modem on the Snapdragon 8 Gen 1. It supports wider 5G bandwidth on the Sub-6Ghz and mmWave networks. As per Qualcomm, the new modem is capable of download speeds of up to 10Gbps. Additionally, it comes with newer features, including uplink carrier aggregation for faster uploads.
On the other hand, Samsung has not named the modem used on the Exynos 2200 but claims that it can support download speeds up to 10Gbps in the E-UTRAN New Radio – Dual Connectivity (EN-DC) mode, which combines LTE and 5G signals for better throughput.
The modems on both Snapdragon 8 Gen 1 and Exynos 2200 support 3GPP’s Release 16 standard for mobile broadband connectivity. Both modems also support Wi-Fi 6 and 6E alongside Bluetooth 5.2.
