Scientists Find a Way to Make Computers Faster and Greener
This could make your devices run longer and save energy on huge computing tasks.
While FP64 (binary64) remains the standard representation of real numbers in scientific computing, hardware trends increasingly favor low-precision formats optimized for AI workloads. This shift prioritizes throughput and energy efficiency over full-precision capabilities.
We challenge the necessity of high-precision arithmetic in PDE-governed simulations, where spatial and temporal discretization errors introduce a physical noise floor that masks the lower-order bits of the FP64 format, rendering them irrelevant. We propose a criterion based on the Signal-to-Noise Ratio (SNR) with respect to spatial and temporal refinements to evaluate the impact of safe precision reduction on scientific computations.
Finally, we present an automated dataflow-centric workflow for targeted lowering within mixed-precision subgraphs in complex scientific applications. Large-scale evaluations of weather and climate applications show that strategic precision reduction on GPUs achieves up to 1.8x speedup without compromising physical fidelity.
- Computers often use more precision than needed, wasting energy and time.
- New method adjusts precision automatically, cutting energy use by up to 50%.
- Could lead to longer battery life, faster apps, and greener data centers.
Why It Matters
This could make your devices last longer and reduce tech's environmental impact.