Services - tools - models - for embedded software development
Embecosm divider strip
Prev  Next

Chapter 5.  Optimizing the GDB Server

5.1. Assessing Performance
5.1.1. Load Generated by Debugging Commands
5.2. Caching SPR and Memory Access
5.3. Compiler Optimization
5.4. Overhead of the RSP Debugger Interface
5.5. Summary of Performance Optimization

As noted earlier, any JTAG action takes many hundreds of system clock cycles. This is due to the serial nature of the protocol and the JTAG clock typically running ten times slower than the system clock.

There are thus two ways to improve performance:

  1. Minimize the number of JTAG actions used.

  2. Maximize the performance of the underlying cycle accurate model.

Embecosm divider strip