Windows 11 will allow you to manage the unified memory of future laptops with NVIDIA RTX Spark

NVIDIA presented its platform RTX Spark the past COMPUTEX, offering users a Windows and Arm alternative in gaming and artificial intelligence laptops. But the platform still doesn’t have a planned release date, although Windows seems to already be preparing to offer some interesting options on laptops with RTX Spark. The guys at Microsoft are working on an option that will allow decide how much unified memory to allocate to tasks.

Windows 11 prepares an option to control the memory of the RTX Spark

The option is called IntelligentCarveoutand is currently hidden within a version of Windows 11 experimentaland will allow you to decide how much memory to allocate for Artificial Intelligence tasks and for graphics tasks. The function allows reserve an amount of memory that would directly be removed from the total memory available for other Windows applications. An option that is sure to be very useful for users who are going to dedicate part of their work with this platform to these tasks.

At the moment, this option is available in the build 29648which belongs to Microsoft’s Experimental Future Platforms channel. This means that it is still being tested at a very early stage, and may even disappear without finally reaching a final version of Windows. Additionally, the configuration of the unified memory is hidden and cannot be accessed through the Windows settings panel.

IntelligentCarveout will allow you to reserve memory for AI and graphics

Even though the RTX Spark are not the only ones that have unified memorywith a large quantity such as 128 GB available on some modelsthis option can be very useful to future users of the platform. In addition, NVIDIA presented these RTX Spark as a portable and powerful solution for Artificial intelligencealthough it also has great performance in graphics tasks.

Microsoft has already modified Windows 11 allowing a upper limit for total accessible system memory. With this increase in memory for the GPU, Windows will allow you to load larger language models on premises or the rendering of larger projects.

Although AMD already has a similar function through its drivers AMD Software: Adrenalinincorporating it directly into Windows can offer a unified system for all platforms that have this type of memorylike the one available in future RTX Spark.