Abstract | ||
---|---|---|
Given the variety of numerical errors that can occur, floating-point programs are difficult to write, test and debug. One common practice employed by developers without an advanced background in numerical analysis is using the highest available precision. While more robust, this can degrade program performance significantly. In this paper we present Precimonious, a dynamic program analysis tool to assist developers in tuning the precision of floating-point programs. Precimonious performs a search on the types of the floating-point program variables trying to lower their precision subject to accuracy constraints and performance goals. Our tool recommends a type instantiation that uses lower precision while producing an accurate enough answer without causing exceptions. We evaluate Precimonious on several widely used functions from the GNU Scientific Library, two NAS Parallel Benchmarks, and three other numerical programs. For most of the programs analyzed, Precimonious reduces precision, which results in performance improvements as high as 41%. |
Year | DOI | Venue |
---|---|---|
2013 | 10.1145/2503210.2503296 | SC |
Keywords | Field | DocType |
floating-point program variable,highest available precision,floating-point program,numerical error,numerical analysis,present precimonious,dynamic program analysis tool,numerical program,lower precision,tuning assistant,precision subject,floating-point precision,floating point arithmetic,dynamic program analysis,program optimization | Program optimization,Mixed precision,Computer science,Floating point,Parallel computing,Program analysis,Numerical analysis,Dynamic program analysis,Extended precision,Debugging | Conference |
ISSN | ISBN | Citations |
2167-4329 | 978-1-4503-2378-9 | 69 |
PageRank | References | Authors |
1.82 | 14 | 10 |
Name | Order | Citations | PageRank |
---|---|---|---|
Cindy Rubio-González | 1 | 187 | 11.88 |
Cuong Nguyen | 2 | 207 | 35.89 |
Hong Diep Nguyen | 3 | 138 | 8.93 |
James Demmel | 4 | 4817 | 551.47 |
William Kahan | 5 | 424 | 102.12 |
Koushik Sen | 6 | 3753 | 155.94 |
David H. Bailey | 7 | 926 | 138.47 |
Costin Iancu | 8 | 529 | 34.78 |
David Hough | 9 | 69 | 2.50 |
Rubio-Gonzalez, C. | 10 | 69 | 1.82 |