| Filename | Latest commit message | Latest commit date |
|---|---|---|
d0639879ea
Reconciles the CPU-decoder improvement line onto current main. main already has the 4 GGUF-file decoders (salvaged via #39) plus its own perf work — energy- autotune kernel selection, fused/batched prefill (~1.7x), NEON ssm scan_row, the DevstralResident Metal driver. This adds the complementary CPU-side wins on top, with no changes to any of that: - gemma4::linear: route the seq>1 path through the NEON+threaded matvec_fast (main still had the scalar triple-loop → ~50x slower full-sequence decode). - lib.rs: qmatmul / qmatmul_expert / qmatmul_ok (matmul f32 activations against raw mmap'd quant blocks) + dequant_parallel (parallel backbone dequant, ~6x load) — pure additions. - granite / lfm2moe / nemotron_gguf / qwen35moe: stay-quantized (weights matmul'd raw on the mmap'd blocks, ~2.4x less RAM, UMA-native) + incremental O(N) generation (KV cache + recurrent/short-conv state) instead of O(N²) re-run. Orthogonal to main's Metal/energy-autotune path: my decoders use main's ssm.rs public primitives unchanged (scan_row is internal); the only shared file is lib.rs (additive). Verified on current main: builds clean, 288 lib tests pass (main's + these), granite oracle greedy " Paris" byte-identical. Co-Authored-By: Claude Opus 4.8 <[email protected]> |
||
| .. | ||
| conformance/v1 | ||
| jouleclaw-rs | ||
| vocab | ||
| LICENSE | ||
| README.md | ||
| ROADMAP.md | ||
| SPEC.md | ||
JouleClaw — Energy-Optimized AI Runtime
The substrate for AI that costs Joules, not tokens.
Capability per Joule, not capability per parameter.
What JouleClaw is
JouleClaw is a pure-Rust, omni-modal AI runtime + harness + agentic layer governed by Transaction Science as an open standard (5th in the open-standards quartet, alongside OpenPay, Smart Byte, EOC, and WAI).
Every operation JouleClaw performs walks an explicit energy-tiered cascade, top to bottom, stopping at the first stage that closes the query:
| Tier | Name | Cost class | What runs here |
|---|---|---|---|
| L0 | Cache | picojoules | Content-addressed cache hits |
| L1 | Lawful | nanojoules | Deterministic, pre-compiled primitive (text/code only) |
| L2 | Embed | sub-millijoules | Matryoshka embeddings + nearest-neighbor + hybrid search |
| L3 | Model | joules | Local SSM / ternary / multimodal / diffusion / etc. |
| L4 | Wire | tens of joules | Remote frontier RPC (escape hatch only) |
L1 (Lawful) is only meaningful for text and code modalities. For image, audio, video and 3D generation, the cascade collapses to L0 / L2 / L3 / L4 — the energy savings come from caching plus dispatch to the cheapest model that meets the quality bar.
Why JouleClaw exists
The mainstream "agentic AI" ecosystem treats inference as the entry point. Every state shift, every tool call, every routing decision runs through a 70-billion-parameter stochastic engine, burning kilojoules to do work that deterministic compute or a hot cache could resolve for microjoules.
JouleClaw inverts the default. Inference is the last resort, not the first. Frozen model weights are less trustworthy than current world state; fresh, provenance-stamped retrieval beats synthesis from training cuts. Every operation is metered in microjoules where the hardware permits it, and in millijoules where it doesn't — the spec is honest about the floor on every platform.
The thermodynamic circuit breaker is the safety mechanism. Set an energy budget; when measured consumption exceeds the budget, the breaker trips and the operation halts. Hallucination loops cannot survive an honest joule ledger.
What's here
| Path | Contents |
|---|---|
SPEC.md |
JouleClaw v1 spec — cascade tiers, energy contract, receipt |
CHARTER.md |
Stewardship pattern (shared across open-standards) |
jouleclaw-rs/ |
Apache-2.0 Rust reference implementation |
spec/ |
Tier specifications, energy provenance, receipt format |
conformance/v1/ |
Signed wire vectors for cross-implementation verification |
Reference implementation
jouleclaw-rs/ is the Apache-2.0 reference implementation in Rust. It
runs on Apple Silicon (M3/M4/M5 with UMA + Metal 4 Tensor APIs), AMD Strix
Halo / Ryzen AI Max, Intel Lunar Lake, NVIDIA Jetson, and discrete
NVIDIA / AMD GPUs.
It is omni-modal: text, code, vision, audio, video, image diffusion, 3D Gaussian splatting, and cross-modal fusion all dispatch through the same energy-tiered cascade.
The MCP tool surface is a JCP consumer: jouleclaw-jcp wraps any
jouleclaw-mcp metered tool so every dispatch is decided by the JCP runtime
against a signed, metered Grant before a byte reaches the tool — the tool's
identity is the content-addressed resource, the call's energy estimate is
debited from the grant's joule budget, and every decision (allow and deny
alike) seals as a JCR-1 receipt. An injected instruction can name an ungranted
tool; it cannot make the gate dispatch it.
Status
v0.1.0 reference implementation in active development. The spec is draft status. Cross-implementation conformance vectors will land with v1.0.0.
License
- Code: Apache-2.0 (see
LICENSE) - Spec text: CC-BY-4.0 (see
spec/LICENSE)