Microsoft: CPU and GPU optimization help Microsoft Teams improve energy efficiency by 50%

Microsoft Teams

Microsoft has just announced that the energy efficiency of the Microsoft Teams service has been significantly improved compared to June 2020, a year and a half ago. In a blog post, Robert Aichner, the company’s chief group program director, mentioned that in the Intelligent Communication and Conversational Cloud (IC3) business, they have achieved up to 50% energy efficiency improvement through CPU /GPU optimization.

Before the evaluation, the team built a suitable framework to test different scenarios such as group video calls and screen sharing.

Microsoft said that this move will not only help save energy and reduce emissions, but also allow the majority of Windows users to get a good experience on low-end hardware.

For enterprises, related optimizations also reduce the burden on their hardware, improve the efficiency of Microsoft Teams meetings, and the consistency of the calling experience across devices.

Clearly, with the rapid adoption of the Teams app, the development team faced some daunting challenges, especially the need to build a consistent experience across the extremely diverse ecosystem of Windows devices.

Microsoft added that we are committed to ensuring a great calling and meeting experience for users with high-end and low-end hardware and high-resolution screens.

Microsoft Teams Performance

Microsoft Teams Performance

Although it cannot be done in one step, Microsoft has gradually improved and tried it in different versions (as shown in the time chart below).

In short, Microsoft Teams has ushered in three major improvements:

  ● Optimize camera functions to reduce CPU workload.

  ● Improve video rescaling when handling group calls of different resolutions.

  ● Improve GPU-based hardware acceleration experience.

On this basis, Microsoft also stated that it will continue to work closely with CPU/GPU suppliers to continuously improve the efficiency of Microsoft Teams and fully prepare for the optimization of next-generation chips.

Leave a Reply