TLDR
Key insights
The spec
THE DELIVERABLE
A dedicated batch-1 decode engine. Roughly 15–30 TOPS INT4-equivalent, a three-tier memory hierarchy with no off-package HBM, and a 5W envelope — targeting 8–15 tok/s on a 100–500B-total heavily-sparse model.
Memory hierarchy
TIER 0
TIER 1
TIER 2
Power budget — 5W
The arithmetic
| Scenario | GB / token | @100 GB/s | @273 GB/s | @614 GB/s | @3 TB/s | @5W energy-limited |
|---|
CROSSOVER
P* = Bandwidth × J/byte
Independent of model size. At 100 GB/s phone memory with a system-level 0.16 nJ/byte,
P* ≈ 16W — below which the device is energy-bound, above which it is
bandwidth-bound. A 5W handheld sits firmly on the energy side. The 5W column is the
binding one.