In addition to performance efficiency, code-size efficiency is an important metric, especially when zero wait-state memory is limited. Performance critical code is usually run out of zero wait-state memory, and non performance critical code is run from Flash memory. Code-size Benchmarks shows code-size efficiency of various benchmarks, comparing C29 with C28 and ARM (Cortex-M7). A few points are noted from the results:

• C29 code-size is mostly comparable to C28, as well as Cortex-M7 code-size. In some benchmarks, the C29 achieves lower code-size, and in some other benchmarks higher code-size.
• Code-size results correspond to the -O3 optimization setting of the compiler. The user has the flexibility to selectively use -Oz on portions of code to reduce code-size.
• C29 FIR code-size is larger because of software pipelining (which results in a huge performance boost). Loops, in general, are a small part of overall code.