In high-level modeling, the overhead of custom transaction classes is a bottleneck. TLM 2.0 introduces the Generic Payload (GP)—a pre-defined, memory-mapped transaction container that serves as the "Universal Language" for architectural models.
Why use the Generic Payload?
- Interoperability: Component A (custom) can talk to Component B (third-party) without class conversion.
- Performance: Optimized for multi-threaded simulation speeds.
- Simplicity: No need to define new classes for standard memory reads/writes.