This session focuses on the compiler’s floating-point arithmetic, which is more complicated to handle than an integer or fixed-point arithmetic. There are many performance and precision glitches that a developer can run into and compilers are not as free to optimize the code automatically. For this reason, a software developer must be more aware of the floating-point arithmetic properties to get good performance from the compiler.
Join compiler architect Andy Kaylor for deeper insights, including:
- Command-line options to control compiler behavior
- Pragmas to fine-tune controls within your code
- How the compiler selects math library function implementations
- The effect of floating-point settings on compiler optimizations
- Trade-offs between performance, accuracy, and reproducibility