Randall Fisher's Home Page at Aggregate.Org

Contact Information:

Career Background:

These were up-to-date as of December, 2012:

Spring Courses at Western Michigan University

Research:

The SWAR parallel processing model:

My research involves an old idea that we are revisiting for use with modern processors, which we call SWAR: SIMD Within A Register. My preliminary Ph.D. exam, General-Purpose SIMD Within A Register : Parallel Processing On Consumer Microprocessors, talks about generalizing this idea (which had been primarily targeted at specific algorithms) into a general-purpose SIMD parallel processing paradigm.

Multimedia extensions use SWAR to speed up algorithms related to multimedia applications, but don't really provide a complete general-purpose model. The goal of this work was to provide such a model, and some basic implementations, to the general public.

I presented an introduction to SWAR and an overview of issues that need to be addressed when designing a SWAR-based computational model (such as a language) at the spring meeting of the Midwest Society for Programming Languages and Systems (MSPLS). This meeting was held at Loyola University Chicaga on May 16, 1998. Here are the slides for the talk: Issues in the Design of SWAR Programming Models.

I also presented a more in-depth description of the requirements for compiling for SWAR targets at the 11th annual Workshop on Languages and Compilers for Parallel Computing (LCPC98). This took place at The University of North Carolina at Chapel Hill, on August 7-9, 1998. Here is a draft of the slides for the talk: Compiling for SIMD Within A Register. This was a presentation of a paper of the same name. The final version (ps, pdf), (© Springer-Verlag), has been published as part of the LNCS series.

On Sept. 23rd, 1998, I gave a talk at the Parallel Processing Seminar at Purdue University, which was a combination of information from the above presentations.

I also presented a more in-depth description of the requirements for compiling for SWAR targets at the 12th annual Workshop on Languages and Compilers for Parallel Computing (LCPC99). This took place at The University of California at San Diego, on August 4-6, 1999. Here are the slides for the talk: The Scc Compiler: SWARing at MMX and 3DNow!. This was a presentation of a paper of the same name. The final version (© Springer-Verlag), will be published as part of the LNCS series.

Here are a few web pages about SWAR:

Parallel Computing Clusters:

I used to work on clusters when I was a Ph.D. student at Purdue. This was before clusters caught on.

Teaching:

Links for Students of Languages, Compilers, and Operating Systems:

Some Parallel Processing Links:



This page was last modified on .