LinkHarvestDigest

Folding Proteins on a MacBook: Bringing Alphafold3 to Apple Silicon Introduction Protein folding has long been one of the most computationally demanding problems in biology. For years, tools like Alphafold were synonymous with high-performance computing (HPC) clusters, GPU farms, and specialized infrastructure. Running these models required access to powerful servers, making them inaccessible to most individual researchers, students, or hobbyists. That landscape is changing. A new side project demonstrates that Alphafold3 can now run smoothly on Apple Silicon, specifically M-series MacBooks released around 2023 and later. This breakthrough means that protein structures can be generated from sequences in minutes—without the need for expensive clusters or cloud credits. What’s New The project, humorously titled “CUDA, Shmuda” , highlights the shift away from NVIDIA’s CUDA-dominated ecosystem. Apple’s M-series chips, with their unified memory architecture and powerful GPU cores, are provi...