feat: add d4r inference engines and Windows parity tooling

Add native K/M/RR inference engines, HIP kernels, replay and validation
harnesses, and Windows DLSS setup and runtime support. Include GPU
profiling tools, build and packaging scripts, platform integration
patches, and implementation documentation.

tools/windows/gpu_targets.py

@@ -0,0 +1,10 @@
+"""Canonical Windows build target registry, shared with CMake and PowerShell."""
+import json
+import pathlib
+
+data = json.loads(pathlib.Path(__file__).with_name('gpu-targets.json').read_text())
+if data['schema'] != 1:
+    raise RuntimeError('Unsupported GPU target registry schema')
+TARGETS = {row['architecture']: row for row in data['targets']}
+if len(TARGETS) != len(data['targets']):
+    raise RuntimeError('Duplicate GPU target')