Vobla

Abstract

We present VOBLA, a domain-specific language designed for programming linear algebra libraries. VOBLA is compiled to PENCIL, a domain independent intermediate language designed for efficient mapping to accelerator architectures such as GPGPUs. PENCIL is compiled to efficient, platform-specific OpenCL code using techniques based on the polyhedral model. This approach addresses both the programmer productivity and performance portability concerns associated with accelerator programming. We demonstrate our approach by using VOBLA to implement a BLAS library. We have evaluated the performance of OpenCL code generated using our compilation flow on ARM Mali, AMD Radeon, and AMD Opteron platforms. The generated code is currently on average 1.9x slower than highly hand-optimized OpenCL code, but on average 8.1x faster than straightforward OpenCL code. Given that the VOBLA coding takes significantly less effort compared to hand-optimizing OpenCL code, we believe our approach leads to improved productivity and performance portability.

Verified code generation for the polyhedral model

“…Another use for the polyhedral model is to synthesize efficient software or hardware implementations of matrix and tensor computations. Domain-specific languages such as Halide [Ragan-Kelley et al 2017], Tensor Comprehensions [Vasilache et al 2020] or VOBLA [Beaugnon et al 2014] make it easy to write high-level specifications of such computations, which can, then, be automatically translated to polyhedral models and compiled to efficient low-level code.…”