{"slug": "nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and", "title": "NVIDIA Details BioNeMo Inference Runtime (BioIR): 2.90x Higher Boltz-2 Folding Throughput and 58.5K Residues per GPU-Hour on 8xH100", "summary": "NVIDIA detailed BioNeMo Inference Runtime (BioIR), a Python library that accelerates biomolecular structure-prediction models on NVIDIA GPUs while remaining in plain PyTorch. In a matched benchmark on 1,000 human dimer targets across 8xH100 GPUs, BioIR-accelerated Boltz-2 delivered 58.5K successfully folded residues per GPU-hour versus 20.2K for a torch-compiled open-source implementation, a 2.90x gain. BioIR already powered the AlphaFold Database expansion, generating about 31 million candidate protein complexes across 4,777 proteomes.", "body_md": "NVIDIA has detailed BioNeMo Inference Runtime (BioIR), a Python library that accelerates biomolecular structure-prediction models on NVIDIA GPUs while staying in plain PyTorch. In a matched benchmark on 1,000 human dimer targets across 8xH100 GPUs, BioIR-accelerated Boltz-2 delivered 58.5K successfully folded residues per GPU-hour versus 20.2K for a torch-compiled open-source implementation, a 2.90x gain. The runtime optimizes at 3 layers: custom kernel selection, CUDA Graph capture, and Ray-based replica scaling that places 1 full model copy per GPU. BioIR already powered the AlphaFold Database expansion, generating about 31 million candidate protein complexes across 4,777 proteomes.\n\nThe post [NVIDIA Details BioNeMo Inference Runtime (BioIR): 2.90x Higher Boltz-2 Folding Throughput and 58.5K Residues per GPU-Hour on 8xH100](https://www.marktechpost.com/2026/09/10/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-throughput-and-58-5k-residues-per-gpu-hour-on-8xh100/) appeared first on [MarkTechPost](https://www.marktechpost.com).", "url": "https://wpnews.pro/news/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and", "canonical_source": "https://www.marktechpost.com/2026/09/10/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-throughput-and-58-5k-residues-per-gpu-hour-on-8xh100/", "published_at": "2026-09-10 21:57:06+00:00", "updated_at": "2026-09-10 22:17:37.657527+00:00", "lang": "en", "topics": ["ai-research", "ai-infrastructure", "machine-learning", "ai-tools"], "entities": ["NVIDIA", "BioNeMo Inference Runtime", "BioIR", "Boltz-2", "PyTorch", "H100", "AlphaFold Database", "Ray"], "alternates": {"html": "https://wpnews.pro/news/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and", "markdown": "https://wpnews.pro/news/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and.md", "text": "https://wpnews.pro/news/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and.txt", "jsonld": "https://wpnews.pro/news/nvidia-details-bionemo-inference-runtime-bioir-2-90x-higher-boltz-2-folding-and.jsonld"}}