新湖畔网 (随信APP) | 高通发布了骁龙 8 至尊版,主频超过 4GHz,Benchmark 多核得分突破了1万。

新湖畔网 (随信APP) | 高通发布了骁龙 8 至尊版,主频超过 4GHz,Benchmark 多核得分突破了1万。
【微信/公众号/视频号/抖音/小红书/快手/bilibili/微博/知乎/今日头条同步报道】

高通的骁龙旗舰芯片近年来的命名变化很多,比如前些年突然出现的骁龙888,目的主要是追求一个吉利。后来变成了骁龙8 Gen1、Gen2、Gen3,到今年骁龙移动芯片的命名开始和PC端命名接近,英文名为Snapdragon 8 Elite,中文名骁龙8至尊版。

取这个名字不仅要有勇气,还需要一些底气,但从这一代骁龙8系列旗舰芯片的表现来看,说是至尊也不过分。

Oryon CPU架构初次进入移动端,全是大核,性能强。以前的移动端芯片多采用big.LITTLE架构,即大小核结合。但从骁龙8 Gen2到骁龙8 Gen3再到今年的骁龙8至尊版,小核数量逐渐减少,甚至直接取消小核,变成全大核架构,2个超大核心和6个大核心。

骁龙8至尊版搭载的新款Oryon CPU架构中,超大核主频高达4.32GHz,而另外6个大核的主频也达到了3.53GHz。这样的架构加持下,在Benchmark 6的测试中,骁龙8至尊版单核得分超过3200,多核超过10000,相较于之前的骁龙8 Gen3有显著提升。

此外,骁龙8至尊版还首次引入Adreno GPU切片架构,使GPU性能提升40%,功耗降低40%。高通展示了在虚幻5引擎中的AI应用和游戏效果,旨在将移动游戏体验提升到主机级别或PC级别。

AI功能也得到升级,Hexagon NPU增加了内核数量,整体推理速度提升45%,支持多个AI模型运行。AI助手在终端侧能更灵敏且无需转换语音为文本。此外,在摄像头辅助下,AI助手也能更好理解周围环境,支持用户个性化需求。

总的来说,骁龙8至尊版在CPU、GPU和AI方面的性能提升明显,致力于提供更高效的多模态AI助手,支持终端侧实时AI处理,如图像和视频处理。#高通骁龙 #至尊版发布主频超 #4GHzBenchmark #多核破 #万

官方原文链接:爱范儿

英文版:

Qualcomm's Snapdragon flagship chips have undergone several changes in naming in recent years, such as the sudden appearance of Snapdragon 888 a few years ago, mainly for luck, and then changed to Snapdragon 8 Gen1, Gen2, Gen3 later. This year, the naming of Snapdragon mobile chips began to approach that of PC chips, with the English name Snapdragon 8 Elite and the Chinese name Snapdragon 8 Supreme Edition.

Choosing this name not only requires courage, but also a little confidence. However, based on the on-paper performance of this generation of Snapdragon 8 series flagship chips, it's fair to say that calling it "Supreme" is not a problem.

Oryon CPU Architecture Debuts on Mobile, All Big Cores, Strong Performance

In the past, many mobile chip designs used the big.LITTLE architecture, which includes big and small cores. This can also lead to a "super big" architecture, where a super big core and big core (performance core) handle heavy tasks, such as large games, while a small core (efficiency core) handles light tasks, such as chatting and web browsing.

However, we can also see that the number of small cores in mobile chips has been decreasing. Snapdragon 8 Gen2 had 3 small cores, which became 2 small cores in Snapdragon 8 Gen3. This year's Snapdragon 8 Supreme Edition simply removed the small cores and became an all-big-core architecture, with 2 super big cores paired with 6 big cores.

Last year, Qualcomm introduced its self-developed Oryon CPU architecture on the Snapdragon X Elite, and this year, the second generation of Qualcomm's Oryon CPU architecture has come to the Snapdragon 8 Supreme Edition. This is also the main reason for the architecture change and performance increase of this flagship chip.

First, the primary frequency of the super big core of the Snapdragon 8 Supreme Edition has reached an astonishing 4.32GHz. Such a frequency appearing on a mobile chip is indeed surprising, as such numbers are usually found on desktop chips.

Even the other 6 big cores have a frequency of up to 3.53GHz.

With this architecture in place, the test model of the Snapdragon 8 Supreme Edition can achieve excellent scores in Benchmark 6, with single-core performance exceeding 3200 and multi-core performance exceeding 10000.

For comparison, the typical scores of the previous-generation Snapdragon 8 Gen3 on the Benchmark official website were 2213 for single-core and 7466 for multi-core.

The Benchmark 6 typical scores of the Snapdragon 8 Gen 2 were 2036 for single-core and 5697 for multi-core.

Therefore, the CPU performance of the Snapdragon 8 Supreme Edition this time is without any issues, according to Qualcomm, the single-core performance and multi-core performance of the Snapdragon 8 Supreme Edition have both increased by 45% compared to the previous generation, while power consumption has also been reduced by over 40%.

In addition to meeting the needs of the AI era, the memory architecture of this generation Oryon CPU has also been greatly upgraded, with 12MB of L2 cache per CPU cluster. Furthermore, the L1 cache of the super big core and big core has also been increased compared to before, and the Snapdragon 8 Supreme Edition also supports the most advanced LPDDR5x memory.

Teamed Up with Unreal Engine 5, GPU Also Enhanced

We know that gaming performance is the result of the combined efforts of CPU and GPU, so this time, the improvement in CPU performance will naturally also enhance gaming performance. An example given by Qualcomm is the Chaos Physics system in the Unreal Engine, which is a simulation tool aimed at empowering the next-generation of games. With the improved CPU performance of the Snapdragon 8 Supreme Edition, it can support over 1000 object collisions and destruction scenes in this simulation tool with a latency of less than 5 milliseconds.

Similar to the CPU architecture update, the Snapdragon 8 Supreme Edition also introduces the Adreno GPU slice architecture for the first time, which increases GPU performance by 40% and reduces power consumption by 40%.

Qualcomm also demonstrated a video running the Nanite solution of Unreal Engine 5, which currently is only supported on the Snapdragon platform. With this feature, theoretically, game developers can render movie-level 3D environments in mobile games.

In the past, we have seen many gaming features of Snapdragon flagship chips, one side is to lower energy consumption to ensure basic image quality experience, such as variable resolutions and intelligent frame interpolation, while the other side is to raise the image quality limit, such as mobile ray tracing, etc.

This time, the strong tie between Qualcomm and Unreal Engine 5 at the conference seems to reveal the ambition to bring mobile game experience closer to console-level or PC-level games. The games like "永劫无间手游版" and "剑侠情缘网络版三无界版" that appeared in the keynote at the conference also tend to make mobile games and PC platform gameplay and data closer, even interconnected.

AI and Imaging, Gradually Integrated, Gradually Strengthened

When Snapdragon 8 Gen3 was released last year, the AI part of the demonstration focused on on-device large language model for text and Stable Diffusion for images, with the source mainly being text.

However, this year, due to the popularity of the AI Agent concept, AI applications have shown a very clear trend towards multimodality, no longer just text-based.

Because AI tasks are very diverse, the AI engine of the Snapdragon 8 Supreme Edition has also been upgraded. The Oryon CPU is mainly responsible for latency-sensitive tasks and the distribution of AI tasks, such as allocating tasks to the sensor hub.

The Hexagon NPU is the central hub of Qualcomm's AI engine, which adds additional cores to the scalar and vector accelerators inside the NPU. Now, the vector accelerator has 6 cores, the scalar accelerator has 8 cores, and the tensor processor has been enhanced, allowing the overall inference speed of the NPU to increase by 45%. To meet the demands of AI multi-tasking, Qualcomm has also enhanced concurrency, allowing AI and computer vision workloads to coexist in NPU memory.

After CPU, GPU, and NPU have all received significant performance improvements, Qualcomm's goal is quite clear: to be able to run multimodal AI assistants on the device, run multiple AI models, including LMM (Multi-Modal Large Model), LLM (Large Language Model), LVM (Visual Large Model), ASR (Automatic Speech Recognition), and more.

For example, in the past, users found that voice assistants were often slow to respond, but now flagship phone voice assistants are more responsive and lifelike, with almost no delay. This is due to technological advancements.

Previously, voice assistants needed to convert voice to text, make the AI model understand the text, organize it into text, process it through a large language model, and then generate the response text, which is then converted back into voice output for the user.

Now, some AI assistants can directly understand the voice without the need to convert it to text, greatly reducing processing time.

In addition to enhanced text and voice comprehension capabilities, AI assistants with the support of the Snapdragon 8 Supreme Edition can now use the camera to directly understand what they see on the device. For example, by pointing the camera at a receipt, the AI assistant can understand the content of the receipt, infer the information, and help calculate payment details, how to split the bill with friends, etc.

In addition to obvious sensors like the camera and microphone, other sensors can also understand the surrounding environment in real time. In combination, based on the user's specific preferences and situations, they can create a user's personal neural network, achieving a more personalized user experience and better insight into user needs.

The ISP (Image Signal Processor) of the Snapdragon 8 Supreme Edition has been enhanced: the new AI ISP works in cooperation with the NPU at the hardware level, greatly increasing pixel throughput to reach 4.3 billion pixels per second. The Snapdragon 8 Supreme Edition supports simultaneous video recording at 30fps with 3 48-megapixel cameras.

Furthermore, because the Hexagon NPU can be directly connected to the memory, this direct connection technology also allows the NPU to directly access the raw RAW data on the ISP, enabling real-time AI-assisted enhancement for 4K 60fps shooting, achieving various AI effects.

By directly introducing AI algorithms into the image and video processing pipeline, various AI functions and effects can be obtained, such as Limitless Segmentation, which supports over 250 layers of semantic recognition and segmentation for images, making it possible to make specific adjustments and improvements to all parts of a photo.

Some practical AI video effects include AI fill light, where even in a heavily backlit environment during a video call, users can add virtual light sources to ensure clear facial images. Another essential feature is AI video magic erasure, where many phones now support AI photo erasure, eliminating unnecessary clutter from photos. In AI video magic erasure, users can select objects to erase in the video, ensuring a clean picture throughout the video, all done on the device with the Snapdragon 8 Supreme Edition, without the need for cloud computing.

 

2024 has shown signs of recovery in the mobile industry, and the concept of AI phones may also reach a turning point. Last year, we saw more of the less usable text and picture demonstrations, while this year we saw more concrete and useful services that are available and needed. All of this coincides with the significant performance increase of the Qualcomm Snapdragon 8 Supreme Edition, the collective design shift of the Android camp, and the early release pace. The fierce competition results in the fact that product capabilities cannot be squeezed further, giving users more choices.

In the plastic greenhouse of fate, in the heart of each cabbage that has been sprayed with too much pesticide, there was once a dream of becoming a pollution-free organic vegetable.

#Follow the official WeChat public account of iFanr: iFanr (WeChat ID: ifanr), bringing you more exciting content first.

iFanr |
Original article link ·
View comments ·
Sina Weibo


高通骁龙 8 至尊版发布,主频超 4GHz,Benchmark 6 多核破 1 万
#高通骁龙 #至尊版发布主频超 #4GHzBenchmark #多核破 #万

关注流程:打开随信App→搜索新湖畔网随信号:973641 →订阅即可!

 

公众号:新湖畔网    抖音:新湖畔网

视频号:新湖畔网    快手:新湖畔网

小红书:新湖畔网    随信:新湖畔网

百家号:新湖畔网    B站:新湖畔网

知乎:新湖畔网      微博:新湖畔网

UC头条:新湖畔网      搜狐号:新湖畔网

趣头条:新湖畔网    虎嗅:新湖畔网

腾讯新闻:新湖畔网    网易号:新湖畔网

36氪:新湖畔网      钛媒体:新湖畔网

今日头条:新湖畔网    西瓜视频:新湖畔网

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注