Automatic translation of FORTRAN programs to vector form
ACM Transactions on Programming Languages and Systems (TOPLAS)
Analysis of interprocedural side effects in a parallel programming environment
Proceedings of the 1st International Conference on Supercomputing
Interprocedural transformations for parallel code generation
Proceedings of the 1991 ACM/IEEE conference on Supercomputing
Interprocedural constant propagation: a study of jump function implementation
PLDI '93 Proceedings of the ACM SIGPLAN 1993 conference on Programming language design and implementation
Effective partial redundancy elimination
PLDI '94 Proceedings of the ACM SIGPLAN 1994 conference on Programming language design and implementation
High-level semantic optimization of numerical codes
ICS '99 Proceedings of the 13th international conference on Supercomputing
An annotation language for optimizing software libraries
Proceedings of the 2nd conference on Domain-specific languages
A case for source-level transformations in MATLAB
Proceedings of the 2nd conference on Domain-specific languages
Evaluating derivatives: principles and techniques of algorithmic differentiation
Evaluating derivatives: principles and techniques of algorithmic differentiation
An Implementation of Interprocedural Bounded Regular Section Analysis
IEEE Transactions on Parallel and Distributed Systems
Otter: Bridging the Gap between MATLAB and ScaLAPACK
HPDC '98 Proceedings of the 7th IEEE International Symposium on High Performance Distributed Computing
IPDPS '00 Proceedings of the 14th International Symposium on Parallel and Distributed Processing
An apl machine
Automatic Type-Driven Library Generation for Telescoping Languages
Proceedings of the 2003 ACM/IEEE conference on Supercomputing
In search of a program generator to implement generic transformations for high-performance computing
Science of Computer Programming - Special issue on the first MetaOCaml workshop 2004
Improving Compilation of Java Scientific Applications
International Journal of High Performance Computing Applications
A language for the compact representation of multiple program versions
LCPC'05 Proceedings of the 18th international conference on Languages and Compilers for Parallel Computing
Combining performance aspects of irregular gauss-seidel via sparse tiling
LCPC'02 Proceedings of the 15th international conference on Languages and Compilers for Parallel Computing
Exploiting domain knowledge to optimize parallel computational mechanics codes
Proceedings of the 27th international ACM conference on International conference on supercomputing
Hi-index | 0.00 |
At Rice University, we have undertaken a project to construct a framework for generating high-level problem solving languages that can achieve high performance on a variety of platforms.The underlying strategy, called telescoping languages, builds problem-solving systems from domain-specific libraries and scripting langauges. To accomplish this it extensively preanalyzes and transforms the library to produce a scripting language precompiler that optimizes library calls within the scripts as if they were primitives in the underlying language.A major technical issue is how to preoptimize a library for use in applications that are yet to be seen. To address this issue, we have conducted a study of applications written in Matlab by the signal processing group at Rice University. This has identified a collection of old and new optimizations that show promise for this particular domain. Two promising new optimizations are procedure vectorization and procedure strength reduction.The latter of these is particularly useful in this problem domain and we expect it to be just as applicable in other contexts as well.We report on the results of an exploration of the effectiveness of procedure strength reduction on three real Digital Signal Processing (DSP) applications. By transforming these programs according to the strategies described in this paper, we were able to achieve speedups ranging from 10 to 40 percent over the entire application — speedups for individual functions were even more dramatic.