ScaleFold: Reducing AlphaFold Initial Training Time to 10 Hours
This work addresses the high computational cost of training protein folding models, enabling faster development and iteration for researchers in computational biology, though it is incremental as it optimizes an existing method rather than introducing a new paradigm.
The paper tackled the problem of AlphaFold2's prohibitively long training time by identifying inefficient communications and overhead-dominated computations as key scaling bottlenecks, and introduced ScaleFold, a systematic training method that reduced the initial training time from seven days to 10 hours and achieved over 6x speedup in benchmarks.
AlphaFold2 has been hailed as a breakthrough in protein folding. It can rapidly predict protein structures with lab-grade accuracy. However, its implementation does not include the necessary training code. OpenFold is the first trainable public reimplementation of AlphaFold. AlphaFold training procedure is prohibitively time-consuming, and gets diminishing benefits from scaling to more compute resources. In this work, we conducted a comprehensive analysis on the AlphaFold training procedure based on Openfold, identified that inefficient communications and overhead-dominated computations were the key factors that prevented the AlphaFold training from effective scaling. We introduced ScaleFold, a systematic training method that incorporated optimizations specifically for these factors. ScaleFold successfully scaled the AlphaFold training to 2080 NVIDIA H100 GPUs with high resource utilization. In the MLPerf HPC v3.0 benchmark, ScaleFold finished the OpenFold benchmark in 7.51 minutes, shown over $6\times$ speedup than the baseline. For training the AlphaFold model from scratch, ScaleFold completed the pretraining in 10 hours, a significant improvement over the seven days required by the original AlphaFold pretraining baseline.