JavaScript is disabled for your browser. Some features of this site may not work without it.
Browsing by Author "Stodghill, Paul"
Now showing items 417 of 17

The Bernoulli Generic Matrix Library
Mateev, Nikolay; Pingali, Keshav; Stodghill, Paul (Cornell University, 20000727)We have implemented the Bernoulli generic programming system for sparse matrix computations. What distinguishes it from existing generic sparse matrix libraries is that we use (i) a highlevel matrix abstraction for writing ... 
Compiling Imperfectlynested Sparse Matrix Codes with Dependences
Ahmed, Nawaaz; Mateev, Nikolay; Pingali, Keshav; Stodghill, Paul (Cornell University, 20000307)We present compiler technology for generating sparse matrix code from (i) dense matrix code and (ii) a description of the indexing structure of the sparse matrices. This technology embeds statement instances into a Cartesian ... 
Compiling Parallel Sparse Code for UserDefined Data Structures
Kotlyar, Vladimir; Pingali, Keshav; Stodghill, Paul (Cornell University, 199706)We describe how various sparse matrix and distribution formats can be handled using the {\em relational} approach to sparse matrix code compilation. This approach allows for the development of compilation techniques that ... 
Dependence Flow Graphs: An Algebraic Approach to Program Dependencies
Pingali, Keshav; Beck, Micah; Johnson, Richard C.; Moudgill, Mayan; Stodghill, Paul (Cornell University, 199009)The topic of intermediate languages for optimizing and parallelizing compilers has received much attention lately. In this paper, we argue that any good representation must have two crucial properties: first, the ... 
A Generic Programming System for Sparse Matrix Computations
Mateev, Nikolay; Kotlyar, Vladimir; Pingali, Keshav; Stodghill, Paul (Cornell University, 199907)Sparse matrices are stored in compressed formats in which zeros are not stored explicitly. Writing highperformance sparse matrix libraries is a difficult and tedious job because there are many compressed formats in use ... 
A Generic Programming System for Sparse Matrix Computations (REVISED)
Nikolay Mateev, Vladimir Kotlyar, Keshav Pingali; Stodghill, Paul (Cornell University, 199908)Sparse matrices are stored in compressed formats in which zeros are not stored explicitly. Writing highperformance sparse matrix libraries is a difficult and tedious job because there are many compressed formats in use ... 
Man vs. Machine : Comparing Handwritten and Compilergenerated ApplicationLevel Checkpointing
Ezick, James; Marques, Daniel; Pingali, Keshav; Stodghill, Paul (Cornell University, 20041012)The contributions of this paper are the following. We describe the implementation of the $C^3$ system for semiautomatic applicationlevel checkpointing of C programs. The system has (i) a precompiler that instruments C ... 
Performance Analysis of the Pipe Problem, a MultiPhysics Simulation Based on Web Services
Stodghill, Paul; Cronin, Rob; Pingali, Keshav; Heber, Gerd (Cornell University, 20040216)The ongoing convergence of grid computing and web services has inspired a number of studies on the use of SOAPbased web services for scientific computing. These studies have exposed several performance problems in ... 
A Relational Approach to the Automatic Generation of Sequential SparseMatrix Codes
Stodghill, Paul (Cornell University, 199707)This thesis presents techniques for automatically generating sparse codes from dense matrix algorithms through a process called \emph{sparse compilation}. We will start by recognizing that sparse computations are ubiquitous ... 
A Relational Approach to the Compilation of Sparse Matrix Programs
Kotlyar, Vladimir; Pingali, Keshav; Stodghill, Paul (Cornell University, 199703)We present a relational algebra based framework for compiling efficient sparse matrix code from dense DOANY loops and a specification of the representation of the sparse matrix. We present experimental data that demonstrates ... 
Solving Alignment using Elementary Linear Algebra
Bau, David; Kodukula, Induprakas; Kotlyar, Vladimir; Pingali, Keshav; Stodghill, Paul (Cornell University, 199501)Data and computation alignment is an important part of compiling sequential programs to architectures with nonuniform memory access times. In this paper, we show that elementary matrix methods can be used to determine ... 
Think Globally, Search Locally
Yotov, Kamen; Pingali, Keshav; Stodghill, Paul (Cornell University, 20041104)A key step in program optimization is the determination of optimal values for code optimization parameters such as cache tile sizes and loop unrolling factors. One approach, which is implemented in most compilers, is to ... 
Unified framework for sparse and dense SPMD code generation(preliminary report)
Kotlyar, Vladimir; Pingali, Keshav; Stodghill, Paul (Cornell University, 199703)We describe a novel approach to sparse {\em and} dense SPMD code generation: we view arrays (sparse and dense) as distributed relations and parallel loop execution as distributed relational query evaluation. This approach ... 
XRay : Automatic Measurement of Hardware Parameters
Yotov, Kamen; Pingali, Keshav; Stodghill, Paul (Cornell University, 20041006)There is growing interest in autonomic, selftuning software that can optimize itself on new platforms, without manual intervention. Optimization requires detailed knowledge of the target platform such as the latency and ...