Unverified50% confidenceBenchmarkExact time
Under the above assumptions, an 8192-token prompt's transfer time is about 30% of prefill time on a 100GbE link, about 15% at 200Gb, and about 7% at 400Gb.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/19/2026
First Seen
Valid until: 1/17/2027
Sources
Related Entities
Related Claims
UnverifiedIf the quantized model weights are 150 GB, reading the weights once at a bandwidth of 273 GB/s takes approximately 0.55 seconds, giving a theoretical single-token generation speed of approximately 1.8 tokens/s (excluding computation latency and cross-node communication overhead).75% similarUnverifiedIn a same-configuration comparison, single-request throughput at 38K improved from 69.38 tokens per second to 86.07, and 8-way total throughput at approximately 4K improved from 400.16 tokens per second to 464.15.72% similarUnverifiedIn a test with a repository of about 12GB of random files, about 192KB flowed to the model channel while about 5.10GiB flowed to the storage channel, split into 73 chunks of about 75MB each.70% similarUnverifiedA 70B-parameter model (140 GB of weights) has a theoretical maximum generation speed of approximately 7 tokens/second at 1 TB/s bandwidth.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1069161API
curl https://kongchang.com/api/v1/knowledge/claims/1069161MCP
get_claim(id=1069161)