Hardware
Run Massive-Scale UMAP in Minutes Using Multiple GPUs—Without Losing Accuracy
NVIDIA cuML and cuVS now support multi‑GPU all‑neighbors kNN graph construction in UMAP, enabling end‑to‑end scaling of the dimensionality reduction algorithm to hundreds of billions of vectors. The i
NVIDIA cuML and cuVS now support multi‑GPU all‑neighbors kNN graph construction in UMAP, enabling end‑to‑end scaling of the dimensionality reduction algorithm to hundreds of billions of vectors. The implementation partitions data into balanced clusters, computes local kNN graphs independently, then merges them into a global graph, avoiding all‑to‑all communication and preserving embedding quality. On the MIRACL and Wiki datasets, eight NVIDIA H100 GPUs achieved up to 74× speed‑up over projected CPU runs—reducing an 870 GB UMAP task to about 8 minutes while maintaining high trustworthiness scores.
Related
- Scaling AI Inference Across Multiple GPUs Using NVIDIA TensorRT with Multi-Device Inference Support
- Scaling Biomolecular Modeling Using Context Parallelism in NVIDIA BioNeMo
Source: NVIDIA Developer | 2026-08-18