Module wasm_cpu 

Prepared CPU execution floor for browser WASM. This backend owns no wgpu objects and remains available when the browser exposes no WebGPU adapter. Qualia’s CPU-WASM LLM backend.

This is a first-party execution floor, not a llama.cpp/wllama binding. It reuses Qualia’s GGUF/P64 index, tokenizer, quantized GEMV and transformer mathematics while keeping WebGPU an optional accelerator.

Structs

CpuWasmEngine

Cold-built, allocation-stable CPU decode plan.

CpuWasmStep

Enums

CpuWasmError

Constants

CPU_WASM_DEFAULT_CONTEXT

Mobile-first default for the independent LLM working set. This is not part of the 42 MiB semantic/SLG Sentinel arena; model inference owns a separate, explicitly sized memory domain.

CPU_WASM_MAX_CONTEXT

Explicit safety cap for one contiguous CPU-WASM KV allocation. This belongs to the LLM memory policy and is unrelated to the semantic Sentinel arena.