| 1. |
- Axelsson, Emil, 1978, et al.
(author)
-
Feldspar: A Domain Specific Language for Digital Signal Processing algorithms
- 2010
-
In: Eighth ACM/IEEE International Conference on Formal Methods and Models for Codesign. - 9781424478859 ; , s. 169-178
-
Conference paper (peer-reviewed)abstract
- A new language, Feldspar, is presented, enabling high-level and platform-independent description of digital signal processing (DSP) algorithms. Feldspar is a pure functional language embedded in Haskell. It offers a high-level dataflow style of programming, as well as a more mathematical style based on vector indices. The key to generating efficient code from such descriptions is a high-level optimization technique called vector fusion. Feldspar is based on a low-level, functional core language which has a relatively small semantic gap to machine-oriented languages like C. The core language serves as the interface to the back-end code generator, which produces C. For very small examples, the generated code performs comparably to hand-written C code when run on a DSP target. While initial results are promising, to achieve good performance on larger examples, issues related to memory access patterns and array copying will have to be addressed.
|
|
| 2. |
- Dévai, Gergely, et al.
(author)
-
Efficient Code Generation from the High-level Domain-specific Language Feldspar for DSPs
- 2010
-
In: ODES-8: 8th Workshop on Optimizations for DSP and Embedded Systems, workshop associated with IEEE/ACM International Symposium on Code Generation and Optimization (CGO).
-
Conference paper (peer-reviewed)abstract
- Software for digital signal processors (DSPs) is traditionally highly hardware-dependent and hence porting it to new processors usually requires significant design effort. In this paper we present Feldspar (Functional Embedded Language for DSP and Parallelism), an embedded, high-level, domain-specific language for DSP algorithm design and the compilation techniques we developed for generating C code from specifications written in Feldspar. While Feldspar allows description of algorithms on specification level, we show that with the right set of abstractions and transformations this high level, functional specification can be transformed into C code that is comparable or better than reference, hand-crafted C language implementations. The Feldspar compiler is highly modular and plugin-based, hence future hardware-specific plugins will enable automatic generation of efficient, hardware-specific code. This approach enables the encapsulation of knowledge of hardware completely in the compiler and thus allows description of algorithms in completely hardware-independent, portable manner.
|
|
