Unverified50% confidenceFactExact time
vLLM v0.27.0 enables the sm_107 target for the NVIDIA Rubin architecture and provides an NVLink all-reduce path on SM107.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
Unverifiedllmfit 通过调用 NVML 获取 NVIDIA GPU 的显存总量、可用量和架构型号，NVML 是 nvidia-smi 背后的同一套 C 语言 API 库70% similarUnverifiedNVIDIA provides native support for nvfp4 in its TensorRT-LLM framework.68% similarUnverifiedvLLM is primarily optimized for NVIDIA GPUs and offers less support for CPU inference and low-end hardware than llama.cpp.67% similarUnverifiedNVIDIA designed a dedicated Tensor Core instruction mma.kind::mxf4 for NVFP4 in the Blackwell architecture (SM 100/103, such as the B100 and B200 series), and adopted an FP8 scaling factor mechanism with one scale per group of 16 elements.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/898656API
curl https://kongchang.com/api/v1/knowledge/claims/898656MCP
get_claim(id=898656)