Unverified50% confidenceTradeoffExact time
NVLink's bandwidth is far lower than the intra-card bandwidth of a single GPU, and cross-card communication creates a bottleneck.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTensor parallelism requires extremely high inter-node bandwidth (typically relying on NVLink or InfiniBand) and is unsuitable for high-latency public network environments; pipeline parallelism vertically partitions by layer, has lower peak bandwidth requirements, and is better suited for heterogeneous network environments.74% similarUnverifiedNVLink Switch的拓扑并非完全扁平，经过不同层级Switch的通信路径在延迟和有效带宽上存在差异71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/907266API
curl https://kongchang.com/api/v1/knowledge/claims/907266MCP
get_claim(id=907266)