TensorRT 11.3 Support Matrix and engine portability rules

TensorRT 11.3.0 support matrix. The load-bearing section for a deployment planner is Engine Portability: 'Serialized engines are not portable across platforms (Linux, Windows, etc.)'; engines built with the version-compatible flag 'can run with newer TensorRT versions within the same major version'; engines built in hardware compatibility mode 'can run on multiple GPU architectures, depending on the hardware compatibility level used. Without this mode, engines are not portable across different GPU architectures.' So a TensorRT engine is an arch-and-version-bound build artifact, not a portable binary - which is the single fact most often assumed wrong before a deployment goes wrong. Also records that Python 3.8/3.9 bindings are deprecated in 11.x and that DLA is not supported in TensorRT 11.3.0/11.3.1 on DriveOS.