No abstract

No abstract

No abstract

We present FGen, a program generator for high performance convolution operations (finite-impulse-response filters). The generator uses an internal mathematical DSL to enable structural optimization at a high level of abstraction. We use FGen as a testbed to demonstrate how to provide modular and extensible support for modern SIMD vector architectures in a DSL-based generator. Specifically, we show how to combine staging and generic programming with type classes to abstract over both the data type (real or complex) and the target architecture (e.g., SSE or AVX) when mapping DSL expressions to C code with explicit vector intrinsics. Benchmarks shows that the generated code is highly competitive with commercial libraries.

This paper describes RandIR, a tool for differential testing of compilers using random instances of a given intermediate representation (IR). RandIR assumes no fixed target language but instead supports extensible IR-definitions through an internal IR-independent representation of operations. This makes it particularly well suited to test embedded compilers for multi-stage programming, which is our main use case. The ideas underlying our work, however, are more generally applicable. RandIR is able to automatically simplify failing instances of a test, a technique commonly referred to as shrinking. This enables testing with large random IR samples, thus increasing the odds of detecting a buggy behavior, while still being able to simplify failing instances to human-readable code.