Earlier today, Microsoft unveiled a comprehensive guide aimed at optimizing internet bandwidth usage for Windows 11 update downloads. This resource delves into various policy settings that administrators can implement to enhance network efficiency within their environments.
Enhancements to Task Manager
For both users and administrators seeking insights into data consumption, the Task Manager remains an invaluable native tool for monitoring hardware performance. With the rise of AI workloads on contemporary PCs, Microsoft has taken steps to enhance Task Manager’s functionality, particularly in relation to AI-focused monitoring capabilities.
A significant update includes the introduction of optional NPU (Neural Processing Unit) and NPU Engine columns across the Processes, Users, and Details pages. This enhancement allows users to gain a clearer understanding of which applications and processes are leveraging the neural processor effectively.
For those unfamiliar with NPUs, these dedicated processors are specifically designed to manage AI and machine-learning tasks more efficiently than traditional CPUs, acting as specialized hardware accelerators. While Windows already provides an overview of overall NPU activity on compatible systems, the new process-level details facilitate a more granular identification of workload sources.
Additionally, Microsoft is incorporating NPU Dedicated Memory and NPU Shared Memory columns into the Details page. These columns offer further visibility into the memory consumption of AI workloads through the NPU, proving particularly beneficial on systems where multiple applications utilize local AI features simultaneously.
The enhancements extend beyond dedicated NPUs; neural engines integrated within GPUs will now also be visible on Task Manager’s Performance page. This development aims to present a more holistic view of AI-related activities across PCs, where acceleration may occur through various hardware components.
Another noteworthy addition is the Isolation column for the Processes and Details pages. These new columns are optional, as Microsoft seeks to maintain a streamlined interface for users who may not require such detailed information. To activate them, users simply need to right-click on one of the column headers in Task Manager and select their preferred metrics.
These new metrics empower users to evaluate how effectively applications are managing AI workloads. Microsoft emphasizes that while applications can be optimized to utilize both the NPU and GPU, not all software takes full advantage of these dedicated resources.
For applications optimized for NPUs, the NPU engine column will indicate either NPU – Compute or NPU – Neural, depending on the processor and driver version. Both designations confirm that on-device AI acceleration is actively engaged. Similarly, applications utilizing the GPU’s dedicated AI hardware will display GPU – Neural in the GPU engine column.
This update also simplifies the identification of applications that may be defaulting to CPU usage. If CPU utilization is high while the NPU remains idle, and the GPU engine does not reflect GPU – Neural, it suggests that the application is not capitalizing on the available dedicated AI hardware. Given that software acceleration via the CPU is inherently slower, such scenarios can lead to overall system sluggishness. Microsoft advises users to further assess the impact by examining the Power usage and Power usage trends columns.
For those interested in exploring the full details, the official blog post is available on Microsoft’s Tech Community website.
