Provided by: INRIA
Date Added: May 2012
Even though media processing applications may rely intensively on floating-point computations, some modern embedded media processors such as the ST231 from the STMicroelectronics ST200 VLIW family do not contain floating-point hardware and provide architectural support only for integer arithmetic. In this paper, the authors present some work in progress on the design and implementation of efficient floating-point software support for embedded integer processors. They provide quantitative evidence of the benefits of supporting various non-generic (that is, specialized, fused, or simultaneous) operations in addition to the five basic arithmetic operations: for individual calls, speedups range from 1.12 to 4.86, while on DSP kernels and benchmarks, their approach allows the users' to be up to 1.34x faster.