docs: FastFlowLM NPU research — AMD NPU runtime, 50 TOPS Strix Halo fit analysis, verified install state, test runbook (2026-08-16)

docs/Dankclaw System Specs.md

@@ -26,6 +26,20 @@ Dankclaw is an AMD APU server for vLLM inference. The host runs Ubuntu 26.04 wit
 | GPU-Visible Memory | 96 GiB (driver VRAM ceiling, overcommitted) |
 | Physical RAM | 30 GiB (shared CPU + GPU) |
 
+### NPU
+
+| Property | Value |
+|----------|-------|
+| Model | AMD XDNA2 (Ryzen AI) |
+| Rating | 50 TOPS |
+| Device | /dev/accel/accel0 |
+| Driver | amdxdna 0.7 (in-tree, kernel 7.0.0-29-generic) |
+| Firmware | 1.1.2.65 |
+| Runtime | FastFlowLM 0.9.45 + XRT (installed 2026-08-16) |
+| Status | Not in service. Research only. See [[FastFlowLM NPU Inference]] |
+
+The NPU shares the 30 GiB unified pool. FLM allocates roughly half of system RAM for NPU use (~15 GiB here). It is idle in production.
+
 ### ROCm
 
 | Property | Value |
@@ -81,3 +95,4 @@ Dankclaw is an AMD APU server for vLLM inference. The host runs Ubuntu 26.04 wit
 - [[Shared Memory Architecture Dankclaw]]
 - [[RAM CPU Offloading in VLLM]]
 - [[Running Qwen3.6 27B FP8 on Dankclaw]]
+- [[FastFlowLM NPU Inference]]