Model Releases
MLCommons Releases New MLPerf Storage v3.0 Benchmark Results
MLCommons announced the release of MLPerf Storage v3.0 benchmark results, expanding the suite to cover the full spectrum of AI storage workloads, including an S3 object‑storage access layer in additio
MLCommons announced the release of MLPerf Storage v3.0 benchmark results, expanding the suite to cover the full spectrum of AI storage workloads, including an S3 object‑storage access layer in addition to the existing POSIX layer. The new version adds two tests—KV Cache for measuring key–value cache performance in LLM inference and Vector Database (VDB) for vector indexing and querying workloads that support high‑dimensional data. These additions provide stakeholders with clearer insights into storage bottlenecks across training checkpointing, KV caching, and vector database use cases.
Related
- Cloud Storage Rapid: Turbocharged object storage for AI and analytics
- LayoutBench: Performance Benchmarking of Cloud Storage Layouts for Multimedia Data
Source: MLCommons | 2026-09-01