Thin and light notebook AI drawing showdown: Ryzen 7 7840S beats i7 13700H, who is stronger?

Notebook products equipped with AMD Ryzen 7040 series mobile processors are already hot-selling in the market. The new generation of Ryzen mobile processors not only adopts the highly efficient Zen4 architecture and leading 4nm process, but also supports local AI hardware acceleration. The Ryzen AI engine is brought to the X86 platform for the first time, which also means that Ryzen notebooks are the first to enter the AI ​​era. In fact, the AI ​​capabilities of Ryzen notebooks do not stop there. With the built-in RDNA3 graphics card of the Ryzen 7040 series mobile processors, it can further provide convenient and fast local AIGC acceleration functions - such as supporting the most popular Stable Diffusion local output. picture.

Ryzen AI RDNA3 GPU, Ryzen 7040 series AI acceleration is more comprehensive

▲Ryzen AI engine provides an extremely efficient local AI acceleration solution for mobile platforms

As mentioned earlier, AMD Ryzen 7040 series mobile processors are the first to have a built-in Ryzen AI engine, which can perform hardware acceleration for local AI applications, thereby providing more energy-efficient local AI solutions for mobile platforms.

In previous articles, we have introduced AMD’s Ryzen AI engine in detail. The AI ​​engine units are interconnected in a dedicated method similar to Mesh. Each unit can communicate directly with each other, so there is no Data blocking situations like traditional CPU architecture also ensure timing certainty. In addition, each AI engine unit is equipped with distributed local memory, which eliminates cache misses. It also has higher access bandwidth and reduces the demand for memory capacity.

Since the built-in AI engine of the Ryzen processor can perform various AI neural network calculations locally, including CNN, RNN, LSTM, etc., without relying on the cloud, it can achieve delay-free processing. In addition, the engine also has real-time multi-tasking capabilities, which can process up to 4 concurrent spatial streams at the same time, and the peak computing power can reach 10TOPS (10 trillion calculations per second)

▲The Ryzen AI engine built into the Ryzen 7040 series mobile processors can achieve local AI acceleration, such as adding background blur, facial tracking, effect enhancement, eye correction and smart beauty special effects to the camera

Currently, many AI calculations are implemented through cloud AI servers and local processors and GPUs. However, now with the Ryzen AI engine exclusive to X86 processors, we can complete these calculations locally with a higher energy efficiency ratio. Especially for notebook computers, the high energy efficiency ratio of the Ryzen AI engine can reduce the AI ​​computing pressure on the processor and GPU, thereby effectively extending battery life. Therefore, even when not plugged in, Ryzen notebooks can utilize the energy-efficient Ryzen AI engine to achieve long-term AI application acceleration

▲The built-in RDNA3 graphics card of Ryzen 7040 can also support Stable Diffusion rendering, and the efficiency is much higher than that of ordinary processors

On the other hand, when users pursue more efficient and broader AI acceleration applications, the RDNA3 graphics card built into the Ryzen 7040 series mobile processors can also flex its muscles. For example, the performance of the Radeon 780M graphics card built into the Ryzen 7 7840S has exceeded that of the GTX 1050 Ti independent graphics card. It can also manually allocate 4GB of video memory. It can run local AI drawing tools such as Stable Diffusion very well and can fully meet the needs of lightweight users. The needs of AIGC users.

Therefore, the Ryzen 7040 series thin and light notebooks, which have both the Ryzen AI engine and the most powerful integrated graphics card, are more versatile in supporting AI applications and can better and more comprehensively meet the needs of daily AI applications and AIGC applications.

Who is the real all-powerful artificial intelligence processor? Ryzen wins overall

Currently, AIGC applications are very popular, among which Stable Diffusion local AI rendering is particularly popular among the public. Compared with online AI drawing, Stable Diffusion has unparalleled freedom and provides a large number of free resources, so the majority of creative design users like to use it. However, in terms of Stable Diffusion local AI rendering acceleration, the GPU has a greater advantage than the CPU. For example, the built-in RDNA3 graphics card of the Ryzen 7040 series can output images faster than conventional processors, and the graphics card itself can also manually set up to 4GB of video memory, which is very useful for increasing the size of AI output images and the number of each batch. help

For users who choose the Ryzen 7040 series thin and light notebook for mobile office, they can use its built-in RDNA3 graphics card to complete some lightweight Stable Diffusion local AI rendering work. This is very practical for digital media workers who often need to create image resources. At the same time, because Stable Diffusion supports DirectML API, it means that the Iris Xe Graphics core display of the Intel platform can also run. So, between the top-of-the-line Iris Xe Graphics core display and Radeon 780M with 96 EU, which one has better AI performance? Next we will conduct actual PK

▲YOGA Air 14s 2023 is equipped with a customized Ryzen 7 7840S processor, built-in Ryzen AI engine and Radeon 780M graphics card

In terms of AMD platform, we chose YOGA Air 14s 2023. Its built-in Ryzen 7 7840S is actually a model specially customized by AMD for Lenovo. It adopts the leading 4nm process and the new Zen4 architecture, and has 8 full specifications. Cores and 16 threads, the highest acceleration frequency is 5.1GHz, and built-in Radeon 780M graphics card. On the YOGA Air 14s 2023, thanks to Lenovo's careful tuning and powerful heat dissipation design, it can achieve a peak power output of 50W in beast mode, which is indeed amazing for an ultra-thin notebook.

In terms of Intel platform, we chose a 14-inch ultra-thin model equipped with Core i7 13700H. Core i7 13700H has a built-in 96 EU Iris Xe Graphics core display, which represents the highest level of Intel core display. In terms of positioning It matches well with Radeon 780M. In terms of memory, both platforms are equipped with 32GB of memory, and the hard drive is PCIe 4.0 SSD 1TB.

In terms of Stable Diffusion, the 4.2 integration package of Akiba aaaki, the UP owner of Bilibili, is used. AMD and Intel platforms use DirectML GPU mode uniformly. The drawing parameters are as follows: number of iteration steps: 20, sampler: DPM 2M Karras, CFG scale: 7, image resolution: 512×512, model: Moyou Artificial Human_V1040, positive prompt words: 1girl, eye contact, sunlight,JK_style.

▲You can choose Radeon 780M and Iris Xe Graphics core display in the Huiyo launcher of the Stable Diffusion integration package

▲The picture output speed of Radeon 780M is about 1 minute and 39 seconds per picture

The time taken by Thunder Dragon 780M to generate 5 batches/5 pictures is 7 minutes and 47 seconds

Both Radeon 780M and Iris Xe Graphics support stable diffusion of AI image generation through DirectML. However, considering that the Radeon 780M is currently the most powerful integrated graphics card, it has a clear advantage in terms of AI computing power. Under our unified parameter settings, it takes about 99 seconds to generate a picture using Radeon 780M, while it takes 192 seconds to use Iris Xe Graphics (96EU). The generation speed is only a little more than half of Radeon 780M

In addition, Radeon 780M can manually allocate 4GB of video memory, so it also has an advantage in continuous drawing. The completion time of 5×1 picture is about 467 seconds, while Iris Xe Graphics (96EU) takes 5×1 continuous drawing in the process of drawing. Stable Diffusion collapsed directly. It can be seen that in the lightweight Stable Diffusion local rendering application, Radeon 780M is not only far more efficient than Iris Xe Graphics (96EU), but its practicality is also more reliable. It can be seen that if you want to use a thin and light notebook to complete the Stable Diffusion local rendering task, then choosing an all-round AI thin and light notebook equipped with a Ryzen 7040 series processor is obviously a much more efficient and practical choice.

The choice of Ruilong for AI notebooks is the real “all-round” summary

In previous tests, we have experienced the AI ​​special effects brought by the built-in Ryzen AI engine of the Ryzen 7040 series mobile processors in video conferencing, and more and more AI applications will also begin to add support for Ryzen AI Engine support. Now through the Stable Diffusion actual AI drawing test, we can also see that the RDNA3 graphics card built into the Ryzen 7040 series mobile processor can also provide excellent local AI acceleration performance for thin and light notebooks. The efficiency of AI drawing is almost that of Intel Iris Xe Graphics (96EU) has twice the core display, and the reliability of continuous drawing is also quite impressive. It has very high practical value for the lightweight AI drawing needs of mobile office users

It can be seen that AMD has taken the lead in AI applications on mobile platforms. The Ryzen 7040 thin and light notebook with both Ryzen AI engine and RDNA3 graphics card can be called the most versatile AI notebook at the moment, which can provide mobile office Users are provided with the most efficient, durable and comprehensive local AI acceleration support, thereby greatly improving work efficiency. All in all, if you want to choose an all-round ultra-thin AI notebook, then products equipped with Ryzen 7040 series processors are undoubtedly the most prioritized solution.

The above is the detailed content of Thin and light notebook AI drawing showdown: Ryzen 7 7840S beats i7 13700H, who is stronger?. For more information, please follow other related articles on the PHP Chinese website!