Unverified50% confidenceFactExact time
310B parameters occupy approximately 620 GB of storage in float16 or bfloat16 precision; completing a full synchronization within 2 seconds implies that the effective bandwidth must reach 310 GB/s or more.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTaking a 70B BF16 model as an example, generating one token requires reading approximately 140 GB of weights, and when fully utilizing the H100's 3.35 TB/s memory bandwidth, the per-token latency is at least 42 milliseconds, with a theoretical TPS upper limit of approximately 24 tokens/second.76% similarUnverifiedTypical home routers are equipped with between 128 MB and 512 MB of RAM, and flash storage usually only 16 MB to 256 MB.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/945487API
curl https://kongchang.com/api/v1/knowledge/claims/945487MCP
get_claim(id=945487)