3D scaling reduces GPU carbon footprint compared to 2.5D scaling
Embodied Emissions of 2.5D and 3D Scaling in Chipletized GPUs for Low-Carbon AI
Hardware Architecture
Summary
As GPUs grow more powerful to support AI, they use more materials that impact the environment. The authors looked at two ways to stack components inside GPUs, called 2.5D and 3D, to see which has a smaller carbon footprint. They found that 3D stacking increases environmental impact less than 2.5D when doubling key parts of the GPU. This means 3D scaling could help make future GPUs greener while still growing in power.
What this means in practice
- •For hardware designers: Choose 3D chip stacking to reduce embodied emissions when scaling GPU compute and memory resources for AI tasks.
- •For data center planners: Incorporate GPUs with 3D scaled chiplets to lower environmental impact of expanding AI infrastructure.
Authors
Hannah Watson, Arjun Singh, Jennifer Stephenson, Job Soethoudt, Debjyoti Bhattacharjee, James Myers
Abstract
GPU compute resources are scaling to support growing AI workloads, leading to further rises in the climate change impact of hyperscale data centers that rely on high-performance GPUs. To model how packaging decisions impact the carbon footprint we have estimated the embodied emissions (EE) of a chiplet-based GPU that has been scaled in 2.5D and 3D using the imec.netzero modelling platform. We doubled the number of compute units and high bandwidth memory (HBM) capacity in the 2.5D and 3D scaled systems and found that 3D scaling results in an 89% increase in EE, as compared to an increase of 141% for 2.5D scaling, positioning 3D scaling as an attractive solution to limit the embodied emissions due to GPU scaling.