Nammu —
Unrolled 8-Step GRU Cooling/Heating Controller
Nammu is the Peabody facility mechroom controller. Named for the Sumerian primeval sea goddess — primordial water from which all creation emerged. Manages the Peabody boiler plant, cooling towers, and chilled/hot water distribution from 34 sensor inputs including tower fan VFD speeds, basin temperatures, condenser water flow, boiler stack temps, and building return water temperatures.
Abstract
Background
Nammu is the Peabody facility mechroom controller. Named for the Sumerian primeval sea goddess — primordial water from which all creation emerged. Manages the Peabody boiler plant, cooling towers, and chilled/hot water distribution from 34 sensor inputs including tower fan VFD speeds, basin temperatures, condenser water flow, boiler stack temps, and building return water temperatures.
Approach
The model was trained in AxonML (a pure-Rust deep learning framework) and compiled through the Hailo Dataflow Compiler (DFC 3.33.1) targeting Hailo-8 silicon. Post-training INT8 quantization was applied during the DFC compilation pass with production telemetry calibration data. The resulting Hailo Executable Format (HEF) binary executes on Hailo’s fixed-function dataflow architecture with deterministic latency and zero framework overhead at the edge.
Results
On production hardware (Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289)), Nammu achieves 59,695 FPS (hw_only) with 0.057 ms hardware latency at 0.94 W average power draw.
Conclusion
Nammu is production-ready as a single HEF binary deployed to edge devices with no external dependencies beyond the HailoRT vendor runtime. The model meets real-time latency requirements for its target hvac prediction application.
| Model | Nammu |
| Domain | HVAC Prediction |
| Architecture | Unrolled 8-Step GRU Cooling/Heating Controller |
| Target silicon | Hailo-8 |
| Measured on | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
| DFC compiler | 3.33.1 |
| Framework | AxonML v0.6 (pure-Rust, CUDA + CPU backends) |
| Author | Andrew Jewell Sr. · ORCID 0009-0005-2158-7060 |
| Organization | AutomataNexus LLC · Fort Wayne, Indiana |
Executive overview
Nammu is the Peabody facility mechroom controller. Named for the Sumerian primeval sea goddess — primordial water from which all creation emerged. Manages the Peabody boiler plant, cooling towers, and chilled/hot water distribution from 34 sensor inputs including tower fan VFD speeds, basin temperatures, condenser water flow, boiler stack temps, and building return water temperatures.
Network I/O
Input: 8-step mechroom telemetry [1, 1, 8, 34→57]. Output: staging/modulation prediction.
Architecture
Unrolled 8-Step GRU Cooling/Heating Controller
8-timestep unrolled GRU with input_size=34 (projected to hidden_size=57). Peabody-specific channels: cooling tower fan speeds (2 cells), basin water temp, condenser water supply/return, chilled water supply/return, boiler supply temps (3 boilers), boiler stack temps, building HW/CHW return temps, outdoor air conditions, pump differential pressures, and electrical demand signal. The model predicts optimal next-period cooling tower fan staging and boiler modulation.
Compilation constraints
All AxonML models targeting Hailo silicon are compiled under the fixed-function dataflow constraints: no dynamic control flow, no variable-length dimensions, all activations representable in INT8 after calibration, and no operations requiring dedicated softmax hardware (replaced with ReLU gating or depthwise convolution equivalents where necessary).
Silicon performance
Measured on production hardware via hailortcli benchmark with 5-second sustained inference. Device: Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289).
| Metric | Measured Value |
|---|---|
| FPS (hw_only) | 59,695.20 |
| FPS (streaming) | 34,744.80 |
| HW Latency | 0.057000 ms |
| Power (streaming avg) | 0.94500 W |
| Power (streaming max) | 0.97200 W |
| Power (idle) | 0.74979 W |
| Quantization | INT8 (post-training, DFC calibration) |
| DFC Compiler | 3.33.1 |
| HailoRT | 4.20.0 |
| Measured On | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
Deployment
Deployed as a single HEF binary. No ONNX runtime, TensorFlow Lite, or Python inference stack required at the edge.
| Target silicon | Hailo-8 |
| Measured on | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
| DFC compiler | 3.33.1 |
| Quantization | INT8 (post-training, production telemetry calibration) |
| Runtime | HailoRT (vendor runtime) |
| Edge platform | Raspberry Pi 5 + Hailo AI HAT+ (M.2 Key M) |
Deployment procedure
Copy the .hef binary to the target device. hailortcli run loads the HEF directly into the Hailo-8 dataflow engine over PCIe. Inference begins immediately with deterministic per-frame latency. No model conversion, graph optimization, or warmup phase required.
References
- Jewell, A. (2026). AxonML: A Pure-Rust Deep Learning Framework for Edge Inference. AutomataNexus LLC. Technical whitepaper.
- Hailo Technologies Ltd. (2024). Hailo Dataflow Compiler User Guide. DFC v3.33.1.
- Hailo Technologies Ltd. (2024). Hailo-8 Product Datasheet.
AutomataNexus LLC · Fort Wayne, Indiana · andrew.jewellsr@automatanexus.com
Andrew Jewell Sr. · ORCID 0009-0005-2158-7060
May 2026 · All rights reserved.