By Ruby Li, Texas Instruments, Inc. Embedded.com Nov 29 2005 (17:55 PM)
Today's intelligent compilers offer many options for squeezing more performance out of application code. Many of these optimizations, however, tend to increase overall code size.
As a result, once developers of optimized application code have reached the required performance specifications, there still remains the challenge of bringing code size back under control.
Through an iterative process of building application code using different compiler optimization options and profiling the result, developers can hone in and identify infrequently used and non-critical sections of code to trade off performance where it matters least for reduced code size, providing minimal impact on system performance. Often, varying compiler options to reduce code size can enable developers to decrease the amount of on-chip and external memory an application requires without adversely affecting performance, thereby reducing the overall bill of materials (BOM).
Click here to read more ...