
These days, in Hawaii, Qualcomm lifted the lid on its newest cell platform — the Snapdragon 8 Gen 2. Construction at the Snapdragon 8 Gen 1 in additional than simply title, Qualcomm’s newest access in its flagship Snapdragon series packs a variety of enhancements and new options for high-end smartphones in 2023 and past.
Whilst you’ll in finding our very important evaluation of the key talking points here, there’s so much to dive into in a lot higher element. Between an all-new CPU cluster association, a ray tracing-capable GPU, high-end audio options and connectivity, and a deeper infusion of imaging and mechanical device studying, the Snapdragon 8 Gen 2 sports activities a large number of firsts for Qualcomm.
See additionally: What is an SoC? Everything you need to know
Snapdragon 8 Gen 2 as opposed to Snapdragon 8 Gen 1
Snapdragon 8 Gen 2 | Snapdragon 8 Gen 1 | Snapdragon 888 | |
---|---|---|---|
CPU Config |
Snapdragon 8 Gen 2
1x 3.19GHz (Cortex-X3) |
Snapdragon 8 Gen 1
1x 3.0GHz (Cortex-X2) |
Snapdragon 888
1x 2.84GHz (Cortex-X1) |
GPU |
Snapdragon 8 Gen 2
Adreno 740 |
Snapdragon 8 Gen 1
Adreno 730 |
Snapdragon 888
Adreno 660 |
DSP |
Snapdragon 8 Gen 2
Hexagon |
Snapdragon 8 Gen 1
Hexagon |
Snapdragon 888
Hexagon 780 |
RAM make stronger |
Snapdragon 8 Gen 2
LPDDR5X |
Snapdragon 8 Gen 1
LPDDR5 |
Snapdragon 888
LPDDR5 |
Digicam make stronger |
Snapdragon 8 Gen 2
• 200MP unmarried shot |
Snapdragon 8 Gen 1
• 200MP unmarried shot |
Snapdragon 888
• 200MP unmarried shot |
Video seize |
Snapdragon 8 Gen 2
8K @ 30fps (HDR) |
Snapdragon 8 Gen 1
8K @ 30fps (HDR) |
Snapdragon 888
8K @ 30fps |
Video playback |
Snapdragon 8 Gen 2
8K as much as 60fps |
Snapdragon 8 Gen 1
8K |
Snapdragon 888
8K |
Charging |
Snapdragon 8 Gen 2
Fast Fee 5 |
Snapdragon 8 Gen 1
Fast Fee 5 |
Snapdragon 888
Fast Fee 5 |
4G/5G Modem |
Snapdragon 8 Gen 2
X70 LTE/5G (built-in) |
Snapdragon 8 Gen 1
X65 LTE/5G (built-in) |
Snapdragon 888
X60 LTE/5G (built-in) |
Different networking |
Snapdragon 8 Gen 2
Bluetooth 5.3 |
Snapdragon 8 Gen 1
Bluetooth 5.2 |
Snapdragon 888
Bluetooth 5.2 |
Procedure |
Snapdragon 8 Gen 2
TSMC 4nm (N4?) |
Snapdragon 8 Gen 1
Samsung 4nm 4LPE |
Snapdragon 888
Samsung 5nm LPE |
Snapdragon 8 Gen 2 CPU association defined

One of the crucial right away noticeable adjustments to the Snapdragon 8 Gen 2 is the transfer from the attempted and examined 1+3+4 CPU cluster association to a extra novel 1+4+3 setup. As well as, Qualcomm has opted for 2 other CPU cores within the heart/efficiency cluster, in response to two more recent Arm Cortex-A715s and two last-gen Cortex-A710s. That’s sure to spice up multi-core benchmarking ratings, nevertheless it’s additionally obviously an overly particular design selection.
In step with Qualcomm, the reasoning got here all the way down to proceeding make stronger for legacy programs. The Cortex-A710 is the final of Arm’s cores to make stronger 32-bit programs (AArch32) — all next and long term cores are 64-bit best (AAarch64), a minimum of in principle. The Snapdragon 8 Gen 2 additionally makes use of Arm’s refreshed Cortex-A510 little cores, which, along side a 5% aid in energy intake, will also be constructed with 32-bit make stronger as of 2022.
Making sure legacy 32-bit make stronger leads to a singular CPU format within the 8 Gen 2.
Qualcomm wasn’t crystal transparent after we requested about AArch32 make stronger within the A510s, however its resolution leads us to imagine that best the 2 A710 cores have 32-bit make stronger for legacy programs. That isn’t essentially very best from a efficiency or potency perspective, however compromises needed to be made someplace, and it is smart to hit legacy moderately than present apps. Even so, 32-bit make stronger may well be redundant for lots of Snapdragon customers and may just also be a deficient trade-off whilst you believe the lack of one little potency core. Then again, Qualcomm claims to have additional optimized the efficiency cores to mitigate the problem.
See, Google has mandated 64-bit software make stronger for years already. Any app up to date at the Play Retailer in recent times is now 64-bit. Even so, together with the A710 cores guarantees that the Snapdragon 8 Gen 2 will paintings with older programs and those who fall out of doors of Google’s Android ecosystem. Assume China or third-party app shops which are additional at the back of in mandating 64-bit make stronger.
An additional heart core will spice up multi-core workloads, however what about low energy use instances?
A powerhouse Arm Cortex-X3 rounds out the CPU clusters, offering a tight chew of the claimed 35% efficiency development, along side the additional heart core. In relation to potency, Qualcomm claims as much as a 40% development total. The majority of this comes from the transfer to TSMC’s 4nm procedure (Qualcomm wouldn’t ascertain whether it is the usage of TSMC’s N4 or more recent N4P procedure, so we’re presuming the previous), nevertheless it’s nonetheless an outstanding determine given the lack of one potency core. We noticed equivalent advantages when Qualcomm moved from Samsung to TSMC for the Snapdragon 8 Plus Gen 1.
Efficiency cluster | Heart Cluster | Potency Cluster | |
---|---|---|---|
CPU cores |
Efficiency cluster
1x Arm Cortex-X3 |
Heart Cluster
2x Arm Cortex-A715 |
Potency Cluster
3x Arm Cortex-A510 |
Clock Pace |
Efficiency cluster
3.19GHz |
Heart Cluster
2.8GHz |
Potency Cluster
2.0GHz |
L1 cache |
Efficiency cluster
(unknown) |
Heart Cluster
(unknown) |
Potency Cluster
(unknown) |
L2 cache |
Efficiency cluster
1MB |
Heart Cluster
(unknown) |
Potency Cluster
(unknown) |
L3 cache |
Efficiency cluster
8MB (shared) |
Heart Cluster
8MB (shared) |
Potency Cluster
8MB (shared) |
64/32-bit make stronger |
Efficiency cluster
64-bit best |
Heart Cluster
2x A715: 64-bit best |
Potency Cluster
64-bit best |
The desk above supplies an outline of the CPU setup, a minimum of so far as Qualcomm would ascertain to us. We don’t have complete cache information, which may have efficiency implications for the center and potency cores. Nonetheless, Qualcomm has equipped a bigger shared L3 cache, now 8MB up from 6MB, which is able to play a task in maximizing efficiency in closely multi-threaded workloads with the extra heart core.
We’ll must watch for smartphones in hand to peer how this shakes out, however some fascinating alternatives had been made right here.
Ray tracing graphics for cell

Now for, arguably, the head-line grabbing characteristic — cell ray tracing graphics {hardware} is going mainstream. Qualcomm isn’t the primary to announce hardware-accelerated ray tracing options for cell; it joins the AMD Xclipse GPU in Samsung’s Exynos 2200 and the Arm Immortalis-G715 inside of Mediatek’s Dimensity 9200. However Qualcomm’s cargo quantity makes this the announcement that may make cell ray-tracing viable for builders.
Frustratingly, Qualcomm helps to keep its Adreno GPU generation a intently guarded secret. However we all know that the Snapdragon 8 Gen 2 speeds up ray-box and ray-triangle intersections. Importantly, there’s Bounding Quantity Hierarchical (BVH) acceleration (or retrieval and decompression acceleration construction nodes, as Qualcomm calls it), too, considerably expanding the GPU’s talent to check ray collisions optimally.
Builders can leverage ray tracing to construct comfortable shadows, reflections, and world illumination into the video games.
In line with those main points, Qualcomm’s implementation turns out extra powerful than different cell ray tracing choices. As an example, Arm’s ray tracing {hardware} does no longer make stronger BVH. Then again, Qualcomm hasn’t instructed us precisely how tough the Snapdragon 8 Gen 2’s accelerator in point of fact is or how smartly its ray tracing {hardware} scales. So we must be skeptical of any claims about graphical constancy or resolutions chances are you’ll go together with console and PC ray tracing functions. Qualcomm’s ray tracing {hardware} runs on sub-5W of energy, in the end.
In step with Qualcomm spouse Oppo, which is bringing the 8 Gen 2 to its subsequent In finding X flagship, the corporate’s open-sourced PhysRay engine is in a position to building up ray tracing rendering potency by means of an element of 5x and cut back CPU workloads by means of 90% when put next with operating the similar results in tool. The corporate claims a locked 60fps at 720p for half-hour, operating its ray tracing engine at the 8 Gen 2.

Nonetheless, the GPU now speeds up the rendering of soft shadows, reflections, ambient occlusion, and world illumination in ray tracing-capable Vulkan Android video games in techniques that may’t be carried out in tool, in keeping with Qualcomm. So video games must glance that bit nicer within the future years. Talking of, Qualcomm sees hardware-accelerated ray tracing arriving within the AAA video games within the first part of 2023.
The primary cell sport with ray tracing make stronger is predicted in H1 2023.
Along with ray tracing make stronger, the newest, anonymous Adreno GPU (we’re having a bet on Adreno 740 internally) guarantees 25% extra efficiency and as much as 45% energy financial savings over the former era, relying at the use case. It helps the Vulkan 1.3 API, and Qualcomm has optimized its drivers to supply an extra 30% efficiency development in some Vulkan-powered eventualities. Qualcomm could also be the primary to assert make stronger for Unreal Engine 5’s Metahumans framework, whilst its Adreno Show engine boasts Adaptive HDR, HDR Brilliant, HDR10+, Dolby Imaginative and prescient, and OLED Getting older Reimbursement options. This all feels like a large win for Snapdragon avid gamers this 12 months.
Additional AI for imaging and extra

Qualcomm has been beautiful scorching on imaging functions in earlier years and, even supposing no longer touting massive numbers for us to ogle presently, does have some a very powerful enhancements right here too. Ahead of attending to the imaging smarts, let’s dive into what’s new with Qualcomm’s newest Hexagon DSP, the center of the Snapdragon 8 Gen 2’s system-wide AI Engine.
A couple of apparently small enhancements upload as much as slightly so much. For starters, there’s now a devoted energy supply formulation, that means that the Hexagon DSP can run with no need to clock different parts concurrently, such because the GPU. A novel energy area is a win for potency. In that vein, Qualcomm claims a 60% development in efficiency in line with watt over the former era when operating sure AI fashions.
Qualcomm’s Hexagon processor doubles its Tensor crunching functions and brings low-resolution INT4 ML make stronger.
To spice up efficiency, the Tensor accelerator throughout the DSP has doubled in measurement for two times the efficiency and has new optimizations in particular for language processing. Qualcomm could also be debuting what it calls micro tile inference make stronger, necessarily slicing up imaging and different issues into smaller tiles to avoid wasting on reminiscence on the expense of a few consequence accuracy. Alongside the ones traces, the addition of INT4 additionally implies that builders can now enforce mechanical device studying issues requiring excessive bandwidth on the expense of a few accuracy if compressing a bigger fashion. Qualcomm is offering equipment to companions to lend a hand make stronger INT4, so it’s going to require a retooling of present programs to paintings.
Altogether, the Snapdragon 8 Gen 2 Hexagon DSP provides 4.35x the efficiency of its predecessor, relying at the ML fashion. (On this case, Qualcomm is evaluating mobileBERT herbal language processing). Sounds spectacular, however I feel the extra vital exchange is the advent of Hexagon Direct Hyperlink, which extra intently connects its ISP to the AI Engine. The corporate dubs this its “Cognitive ISP.”

Qualcomm doubled the bodily hyperlink between the picture sign processor (ISP), Hexagon DSP, and Adreno GPU, using upper bandwidth and decreasing latency. This permits the Snapdragon 8 Gen 2 to run a lot more tough machine-learning duties on imaging knowledge proper off the digicam sensor. RAW knowledge, for example, will also be handed without delay to the DSP/AI Engine for imaging workloads, or Qualcomm can use the hyperlink to upscale low-res gaming eventualities to lend a hand with GPU load balancing.
Hexagon Direct Hyperlink will increase bandwidth for passing imaging and different knowledge without delay to the AI Engine, bypassing slow DDR reminiscence.
Qualcomm’s top use case for Hexagon Direct Hyperlink is for symbol segmentation and processing. In different phrases, figuring out key facets of a scene, reminiscent of facial landmarks, crops, the sky, and so forth., to create layers in real-time after which observe customized processing to those layers ahead of even hitting the shutter button.
If this sounds slightly acquainted, it’s as a result of Qualcomm has been transferring quite a lot of mechanical device learning-type options nearer to the ISP in earlier years, together with face detection and segmentation for video bokeh functions. It undoubtedly claimed segmentation functions last-gen. Then again, the slower hyperlink supposed that imaging knowledge was once in the past regularly pulled into primary reminiscence first, a expensive and high-latency process that normally led to making use of segmentation after seize. Qualcomm is lowering that bottleneck this 12 months, making it a lot more possible to run complicated workloads, reminiscent of imaging issues, on its AI Engine in genuine time. Even supposing whether or not Qualcomm’s product companions leverage those functions is still noticed.
Extra wi-fi connectivity choices

Let’s get started with the extra thrilling, user-defining new connectivity options. The chip’s up to date Snapdragon Sound audio suite now contains Dynamic Spatial Audio functions. By way of dynamic, Qualcomm approach you’ll now transfer your head round inside the area and listen to the content material transfer round you moderately than following your head statically, because of dynamic head monitoring on suitable headphones. This generation works with maximum present multi-channel spatial audio codecs and decoders, reminiscent of Dolby Atmos and Sony 360 Fact Audio.
Sticking with audio, Qualcomm’s aptX Lossless codec is now supported in each Bluetooth Vintage and LE Audio use instances, combining the advantages of low power and lossless audio playback for long term merchandise. For avid gamers, wi-fi latency can fall to simply 48ms with a suitable headset — that’s 47% not up to its predecessor.
Song enthusiasts have the benefit of Dynamic Spatial Audio and lossless Bluetooth streaming.
For those who concept the mud had settled on 5G, assume once more. Qualcomm is shaking a couple of issues up with the Snapdragon 8 Gen 2. Constructed with an built-in Snapdragon X70 modem, which provides speeds of 10Gbps down and three.5Gbps up by the use of 4x service aggregation, there are AI smarts onboard too.
Qualcomm claims that the modem’s AI functions permit it to enhance the throughput and connectivity robustness of each sub-6GHz and mmWave connections, specifically on the mobile edge. In all probability more effective, even though, is make stronger for twin lively 5G SIMs. So you’ll proceed to obtain messages and information on a secondary 5G SIM whilst taking a decision at the first.
Qualcomm rounds out its newest Snapdragon Attach suite with early make stronger for Wi-Fi 7, in addition to Wi-Fi 6 and 6E. Even supposing the spec isn’t finalized, Qualcomm is leveraging its inside of observe to make stronger the usual early. The promise is as much as 5.8Gbps knowledge speeds over a 320MHz channel within the 6GHz band by the use of Prime Band Simultaneous Multi-Hyperlink. This comes with a latency of simply 2ms, which Qualcomm says shall be precious in supporting cloud gaming, XR, and different latency-dependent programs. After all, you’ll desire a Wi-Fi 7 router to profit, however the ones are best on sale in China on the time of writing. One to slap at the futureproof pieces checklist, then.
Different Snapdragon 8 Gen 2 options
Sifting throughout the release displays and press fabrics, listed here are a couple of different Snapdragon 8 Gen 2 options price highlighting:
- That is Qualcomm’s first processor to make stronger AV1 playback, at as much as 8K 60fps. All primary SoCs heading to long term Android telephones now make stronger AV1 interpreting.
- Twin Bluetooth radios promise to double the connectivity vary and accelerate tool pairing.
- Snapdragon 8 Gen 2 is tuned to make stronger new symbol sensors, particularly the 200MP Samsung ISOCELL HP3 with real-time remoasiac, and Sony’s quad virtual overlap HDR video generation within the IMX800 and IMX989.
- Qualcomm hasn’t made spec adjustments to its ISP functions for the reason that 8 Gen 1. There’s the similar 200MP single-shot digicam, 36MP triple digicam seize, and 4K HDR simultaneous seize options as final 12 months.
- Qualcomm has added a 2d AI processor to its 4th era Sensing Hub. Blended with 50% extra reminiscence, there’s now two times the efficiency on be offering right here to leverage applied sciences like Qualcomm’s always-sensing digicam to use privateness display screen options.
Which smartphones come with the Snapdragon 8 Gen 2?

With the SoC announcement best simply scorching off the click, we most probably received’t see the primary smartphones in maximum shoppers’ palms till 2023 — even though Qualcomm notes we would possibly see a industrial free up ahead of the top of 2022.
Qualcomm has showed that the next companions are operating on Snapdragon 8 Gen 2-powered smartphones: Asus, Honor, Iqoo, Motorola, Meizu, Nubia, Oneplus, Oppo, Redmagic, Redmi, Sharp, Sony, Vivo, Xiaomi, and ZTE. Samsung isn’t on Qualcomm’s checklist, as standard, however the Snapdragon 8 Gen 2 shall be powering the Galaxy S23 series too.