At 23:00 on the evening of March 30, Intel officially released the A series of discrete graphics product lines, including the Three Series of Sharp 3, Sharp 5, and Sharp 7. The first laptops with the Ruixuan 3 series have been launched, followed by the Ruixuan 5 series and 7 series products expected to be available this summer.
According to Intel's official positioning, Ruixuan 3 series is oriented to mainstream games, mainly used in thin and light platforms, and Ruixuan 5 and Ruixuan 7 series are positioned as performance games and hardcore performance games respectively, and will be oriented to high-end game platforms. They're all built on the Xe HPG architecture, an architecture designed for gamers and creators that provides a powerful AI engine and an enhanced media engine that supports next-generation codec standards. In addition, Intel has created the next-generation Xe display engine and a new graphics pipeline to handle a variety of different display tasks.
Because it is based on the Xe HPG architecture, all Intel Iris Discrete Graphics will include a common architecture and advanced feature set, including support for the DX12 Ultimate Ultimate Edition and a state-of-the-art AI engine as well as a media engine.
The biggest feature of the Xe HPG architecture is flexibility, and rendering slices are the basic building blocks of reusable IP. Specifically, in the Xe-HPG architecture, every 4 Xe cores form a render slice, and each Xe core is equipped with a considerable number of computing units, such as the vector engine XVE and the matrix engine XMX. In addition, Xe-HPG also integrates other mainstream graphics technologies, such as mesh shading, sampler feedback, etc., which can find their corresponding hardware support in the rendering tiles.
Intel builds different SOCs by overlaying rendering slices, with a minimum of two and a maximum of up to eight, and depending on the number of rendered slices, the product line becomes richer, giving users more choice. Xe HPG's render slices also support the DX12 Ultimate, which includes improvements to all graphics fixing function blocks, and of course, dedicated hardware light tracing units that support Microsoft DXR and Vulkan RT. Each render slice is also equipped with four hardware light tracking accelerators to support real-time ray tracing technology, which can significantly improve the game graphics and light and shadow effects of AAA masterpieces.
Unlike the EU concept of the previous Intel HD, the Xe core (Xe-Core) has become the most basic execution unit in the Xe-HPG architecture. Each Xe-Core includes 16 256-bit wide SIMD vector engines that perform most of the operations for traditional graphics shaders, primarily for the computational tasks of traditional image processing.
Considering that the algorithmic core of AI revolves almost entirely around a series of large matrix multiplication and accumulation algorithms, Intel has also built 16 dedicated matrix engines in each Xe core, each 1024 bits wide, which are used to accelerate AI hardware.
To meet the high bandwidth requirements of matrix, vector, and ray-traced units, each Xe core has a large local memory of 192KB. It can be dynamically allocated between L1 cache and shared local memory (SLM) based on the needs of each workload.
A major application of the matrix engine is the use of AI in real-time rendering, and Intel has also developed the XeSS algorithm, a supersampling technology that delivers higher performance in games compared to traditional high-resolution rendering. XeSS uses neural networks to assist motion vectors to produce beautiful, high-resolution images from low-resolution renderings. Intel said that more games will provide XeSS support adaptation in the future.
Like the Sharp Torch, the Xe Monochrome also offers a powerful Xe Media Engine dedicated hardware acceleration unit with the industry's widest range of encoders and decoders, including H.265/HEVC, H.264/MPEG-4/AVC, VP9, and the industry's first hardware acceleration support for AV1 encoding and decoding.
Compared to traditional codecs/decoders, AV1 offers a better compression ratio and better visuals. The AV1 encoding hardware acceleration in Ruixuan Unique Display is significantly faster than that of traditional software implementations. Currently, FFMPEG, Handbrake, Adobe Premiere Pro, DaVinci Resolve, and XSplit have all integrated support for the AV1 codec.
The Xe display engine is built for today's and tomorrow's display technologies, offering 4 display channels, 1 of which is optimized for low power consumption. Supports HDMI 2.0b, DP1.4a, and future DP 2.0 10G. Meet the needs of high-resolution and high-refresh rate displays, with support for 1440p@360Hz or 4 4k@120Hz HDRs and 2 8K 60Hz HDRs.
Ruixuan unique display has two different chip designs, the larger chip codenamed ACM-G10 (previously rumored DG2-512EU), containing up to 32 Xe cores and optical tracking units, 16MB L2 cache, 256-bit GDDR6 memory; the smaller chip codenamed ACM-G11 (previously rumored DG2-128EU), containing 8 Xe cores and optical tracking units, 4MBL2 cache, 96-bit GDDR6 memory Both chip designs include two Xe multifunction codec engines and a 4-channel display output engine.
In terms of specific parameters, it includes Ruixuan A350M, Ruixuan A370M, Ruixuan A550M, Ruixuan A730M and Ruixuan A770M. The first two products belonging to the Rui Xuan 3 series are now available, while the Rui Xuan 5 and Rui Xuan 7 series will not wait until early summer.
For the specific performance part, Intel has also made a certain display of the Iris 3 series discrete graphics. Compared with the Core i7-1280P processor integrated Rui Torch Xe core display, the Rui Xuan A370M unique game performance improvement is significant, and it can smoothly run most popular games under medium special effects.
In terms of content creation efficiency, after the enhancement of ruixuan A370M alone, it can be increased by 1.3-2.4 times in different application scenarios.
Intel said that the performance improvement in many scenarios does not only come from discrete graphics, but also from the blessing of Deep Link technology based on the entire system. Deep Link is actually an umbrella term that covers different technologies such as dynamic power sharing, supercoding, and super computing power.
In the future, thin and light copies that meet Intel Evo certification can also be paired with Sharp Discrete graphics, and the two work together to provide a better mobile experience.
Intel said that RuiXuan unique display has been supported by many OEM, ODM partners, the future from thin and light to high-performance game books many new products will use Ruixuan graphics card, from the second quarter with Ruixuan 3 series of notebook products are expected to be available, the price from $899, and equipped with Ruixuan 5 and Ruixuan 7 series of games are expected to be released this summer.