Tue 25 Jun 2019 14:20 - 14:40 at 229AB - Static Analysis Chair(s): Martin Rinard

While the cost of computation is an easy to understand local property, the cost of data movement on cached architectures depends on global state, does not compose, and is hard to predict. As a result, programmers often fail to consider the cost of data movement. Existing cache models and simulators provide the missing information but are computationally expensive. We present a lightweight cache model for fully associative caches with least recently used (LRU) replacement policy that gives fast and accurate results. We count the cache misses without explicit enumeration of all memory accesses by using symbolic counting techniques twice: 1) to derive the stack distance for each memory access and 2) to count the memory accesses with stack distance larger than the cache size. While this technique seems infeasible in theory, due to non-linearities after the first round of counting, we show that the counting problems are sufficiently linear in practice. Our cache model often computes the results within seconds and contrary to simulation the execution time is mostly problem size independent. Our evaluation measures modeling errors below 0.6% on real hardware. By providing accurate data placement information we enable memory hierarchy aware software development.

Tue 25 Jun

pldi-2019-papers
14:00 - 15:30: PLDI Research Papers - Static Analysis at 229AB
Chair(s): Martin RinardMassachusetts Institute of Technology
pldi-2019-papers14:00 - 14:20
Talk
Meng WuVirginia Tech, Chao WangUSC
Media Attached
pldi-2019-papers14:20 - 14:40
Talk
Tobias GysiETH Zurich, Switzerland, Tobias GrosserETH Zurich, Laurin BrandnerETH Zurich, Switzerland, Torsten HoeflerETH Zurich
Media Attached
pldi-2019-papers14:40 - 15:00
Talk
Laith SakkaPurdue University, Kirshanthan SundararajahPurdue University, Ryan R. NewtonIndiana University, Milind KulkarniPurdue University
Media Attached
pldi-2019-papers15:00 - 15:20
Talk
Phúc C. NguyễnUniversity of Maryland, Thomas GilrayUniversity of Maryland, Sam Tobin-HochstadtIndiana University, David Van HornUniversity of Maryland, USA
Media Attached