Recently, Intel officially released its first products to officially return to the exclusive display market in many years - the entry-level mobile versions of the Arc A350M, Arc A370M, and its performance can probably be compared with the full-blood version of the GTX 1650.
As a newly listed graphics card, it is very important for the optimization of the software, and Intel has officially done compatibility testing and optimization for many games and applications on the market, in addition to a special software is the graphics card running sub-software 3DMark.
According to Intel, their specific optimizations for Arc graphics allow them to improve performance by 15% in 3DMark's Time Spy and Port Royal tests. However, 3DMark software developer UL does not currently recognize the optimization, if the user uses this optimized version of the driver score will not be recognized by the UL database.
For this reason, Intel also plans to add an optional switch to the April driver update, players can turn on or off special optimizations for 3DMark according to their own needs.
In fact, many people said that the specific optimization of the running score software can only make the score higher, and the actual use does not allow you to get better performance, which can only be regarded as a psychological comfort for the user.
However, the following optimization does have an improvement in performance.
Previously, Intel proposed the Deep Link concept when releasing Xe Max monosymography, allowing Intel CPU HDUs to handle one task at the same time as Intel monochromatic display to achieve higher performance. It has three characteristics: dynamic power sharing, super coding and super computing power.
It is understood that the notebook equipped with the 12th generation Core + Sharp Display can start the dynamic power sharing function, increase the power consumption ratio of the GPU when the GPU utilization rate is high, and increase the power consumption ratio of the CPU when the CPU utilization is high, so as to improve the performance without the same total power consumption.
Intel says dynamic power sharing intelligently and automatically prioritizes engine performance based on workload conditions and allocates more power to Intel Core processors or Intel IRis Graphics as needed, improving the performance of creative and compute-intensive applications by 30 percent.
Super Encoding, on the other hand, uses all available media encoding engines on the CPU and GPU. According to Intel's official data, Super Encoding enables media engines across the platform to be combined to increase video encoding speed by 60% compared to Intel Iris Xe graphics alone.
Finally, there's the Super Computing Power feature, which works like Supercoding and enables calls to multiple compute and AI engines across the Intel platform, including Intel Core processors, Iris Xe Graphics, and Intel Iris Discrete Graphics to accelerate a variety of workloads. Intel's official data shows that with the super computing power, the performance increase can reach 24%.