Intel has announced more detailed information about its self-developed Xe MAX dedicated GPU, as well as its new platform technology: Intel Deep Link. Deep linking basically allows Intel to simultaneously utilize integrated graphics, dedicated graphics, and processors to handle workloads when possible. The combination of Xe MAX and iGPU defeated the NVIDIA MX 350 in terms of graphics workloads, and beat the NVIDIA MX 350 in terms of AI-related workloads.
Intel Xe MAX with Deep Link: Slightly better than MX350 in games, using Deep Link eliminates RTX 2080 in AI-related workloads
The Intel Iris Xe MAX discrete GPU has the same 96 EUs as its integrated counterparts, but unlike the latter, it has a higher TDP package, can increase the clock speed to 1.65GHz, and also contains a dedicated 4GB memory. It has variable rate shadows and Intel DLBoost, and can run on PCIe 4.0. Like its integrated products at the same level, the core architecture is Intel Xe LP.
Intel Iris Xe (iGPU) and Xe MAX (dGPU) are actually the same GPU, so it is not surprising that Intel hopes to find a platform technology that will enable them to concentrate resources and beat the competition. Intel Deep Link is the name of the technology. This is a platform innovation that allows Intel not only to put integrated GPU and dedicated GPU together, but also to put the CPU itself together. Compared to cross-vendor configurations (such as NVIDIA-Intel), only one device can run a specific workload.
“We set out to redefine the role of discrete graphics in thin and light notebook computers, and to meet the needs of more and more creators who want more portability. Iris Xe MAX graphics and Intel® Deep Link Technology are Intel’s plans to implement The extended Xe roadmap is an example of the kind of platform-level innovation that is planned to be brought to market in the future.”
–Roger Chandler, Vice President and General Manager of XPU Products and Solutions, Intel Architecture, Graphics and Software Division
The role of deep link: Deep link aggregates multiple processing engines through a common software framework, thereby bringing new functions and better performance to the PC. The framework helps unleash the creativity of lightweight laptops by maximizing CPU performance, improving artificial intelligence (AI) creation performance, and taking industry-leading coding to a new level.
- The working principle of deep linking: Deep linking technology integrates the processing engine under a common software framework, so that software developers can significantly improve the performance of content creation workloads. Applications can scale certain workloads across integrated and discrete graphics. Examples include:
An additional AI function that can perform inference and rendering on two GPUs to accelerate content creation workloads.
- The industry-leading encoding engine is combined in each GPU through hyper-encoding, so that you can spare time to render the video for viewing or sharing.
The first applications to enable deep linking use Intel® Media SDK, Intel® OpenVINO™ Release Toolkit and Intel® VTune™ Profiler to access the full functionality of the platform. Developers can also use Intel’s powerful cross-architecture oneAPI toolkit to take advantage of Deep Link technology. Developers interested in accessing Deep Link technology can register for future updates and information.
Intel Iris Xe MAX GPU gaming and AI benchmark test
Without some first-party benchmarks, no announcements will be complete, and Intel will not be disappointed.
In the 8/10 scenario given in the benchmark test, Intel Iris Xe MAX notebook easily beat NVIDIA’s MX350.I also appreciate the fact that Intel tried Do not Pick benchmarks and give some examples to illustrate that NVIDIA MX350 beats Intel Xe MAX solutions such as Borderlands 3 and Gears Tactics. However, Intel Iris Xe MAX GPU can provide more consistent performance throughout the game layer and completely destroy the MX350 in heavy AAA games such as Metro Exodus. Since it can reach 30+ fps in most AAA games in a medium-sized environment, I can say that this is the first entry-level gaming solution introduced by Intel in years.
On the other hand, workloads related to AI and productivity are completely different. Since Intel’s Iris Xe MAX GPU already has the DL Boost function, and on the basis of Deep Link, the notebook can also distribute AI workloads to all three devices, so a notebook equipped with Xe MAX can definitely reduce its weight level. It can even beat NVIDIA’s RTX 2080 flagship mobile GPU.In terms of coding, Intel’s benchmark test found that the solution based on Intel’s deep link is a Loud 78% Better than NVIDIA RTX 2080!
Intel’s Deep Link technology is a nerd’s dream. There is almost no waste in the full utilization of available resources. This is exactly how things should be configured, reflecting the concept of “metallization” that makes the console so powerful. The best part is that since Intel was banned, we may actually enable “deep linking” in some gaming applications, but mentioned that they are working to add more applications to the platform.
Dynamic power sharing also allows companies to fine-tune performance when TDP is limited (similar to AMD). From now on, Intel® Iris® Xe MAX graphics products for Acer Swift 3x, Asus VivoBook Flip TP470 and Dell Inspiron 15 7000 2-in-1 will be available. These are the first devices to utilize the 11th generation Intel Core mobile processors, Intel Iris Xe MAX graphics chips and Intel Deep Link technology.
The complete press release issued by Intel Iris Xe MAX is given below: