Files
llmgw/internal/web
Ilker ManapandClaude Opus 5.5 f2c6f12cfb
CI / test (push) Successful in 9m1s
How fast a model would answer on the cards, as this cluster would lay it out
For each layout modelplan.Options offers here — one card, tensor parallel
across a node's cards, llama.cpp's layer split, a pipeline across nodes on
a fabric — modelplan.GPUTokensPerSecond gives one reply's speed: the bytes
a token reads over each card's bandwidth, by its model name (a table, SXM
and PCIe apart), plus two all-reduces a layer for tensor parallel (8 us on
NVLink, 25 us and up on PCIe), turns between cards for a layer split, and
a hop per node over InfiniBand or RoCE for a pipeline. The explorer and a
repository's files show the fastest, with the layout; mixed classes are
judged class by class, FP8 on 8.9 cards only.

Calibrated on cuda01's Quadro RTX 6000 (672 GB/s): qwen3-vl-8b Q8, 8.7 GB,
62.1 tok/s measured, 59 estimated; Qwen3-Coder-30B-A3B Q4_K_M, 18.6 GB,
144 measured, which says a model of experts reads 1.6 times its active
share (ActiveShare now does, on the CPU too). Tensor parallel and fabrics
are not measured here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 00:19:36 +01:00
..
2026-09-29 21:06:42 +02:00
2026-10-21 09:22:42 +02:00
2026-10-21 09:22:42 +02:00