Floating-Point (FP) units in processors are generally limited to supporting a subset of formats defined by the IEEE 754 standard, along with a few target-specific ones (X86 with an 80-bit FP format, and PowerPC performing 128-bit FP arithmetic). As a result, high-efficiency languages and optimizing compilers for high-performance computing are also limited by the FP types supported by these units. However, the pursuit of efficiency and stability on applications has led researchers to investigate a finer control of exponent and fraction bits for finding the right balance between accurate results and execution time and/or energy consumed. For example, numerical computations often involve iterative solvers where the residual error is a function of the ...