SC Conference - Activity Details
Automating the Generation of Composed Linear Algebra Kernels
Authors:| Geoffrey Belter (University of Colorado) |
| E. R. Jessup (University of Colorado) |
| Ian Karlin (University of Colorado) |
| Jeremy G. Siek (University of Colorado) |
| Autotuning and Compilers |
| Tuesday, 02:30PM - 03:00PM |
| Room PB256 |
Memory bandwidth limits the performance of important kernels in many
scientific applications. Such applications often use sequences of
Basic Linear Algebra Subprograms (BLAS), and highly efficient
implementations of those routines enable scientists to achieve high
performance at little cost. However, tuning the BLAS in isolation
misses opportunities for memory optimization that result from
composing multiple subprograms. Because it is not practical to create
a library of all BLAS combinations, we have developed a
domain-specific compiler that generates them on demand. In this
paper, we describe a novel algorithm for compiling linear algebra
kernels and searching for the best combination of optimization
choices. We also present a new hybrid analytic/empirical method for
quickly evaluating the profitability of each optimization. We report
experimental results showing speedups of up to 130% relative to the
GotoBLAS on an AMD Opteron and up to 137% relative to MKL on an Intel
Core 2.
The full paper can be found in the
ACM Digital Library and IEEE Computer Society
   |