Tue 25 Jun 2019 16:00 - 16:20 at 224AB - Performance Chair(s): Ting Cao

Minimizing cache misses has been the traditional goal in optimizing cache performance using compiler based techniques. However, continuously increasing dataset sizes combined with large numbers of cache banks and memory banks connected using on-chip networks in emerging manycores/accelerators makes cache hit–miss latency optimization as important as cache miss rate minimization. In this paper, we propose compiler support that optimizes both the latencies of last-level cache (LLC) hits and the latencies of LLC misses. Our approach tries to achieve this goal by improving the parallelism exhibited by LLC hits and LLC misses. More specifically, it tries to maximize both cache-level parallelism (CLP) and memory-level parallelism (MLP). This paper presents different incarnations of our approach, and evaluates them using a set of 12 multithreaded applications. Our results indicate that (i) optimizing MLP first and CLP later brings, on average, 11.31% performance improvement over an approach that already minimizes the number of LLC misses, and (ii) optimizing CLP first and MLP later brings 9.43% performance improvement. In comparison, balancing MLP and CLP brings 17.32% performance improvement on average.

Tue 25 Jun

Displayed time zone: Tijuana, Baja California change

16:00 - 17:00
PerformancePLDI Research Papers at 224AB
Chair(s): Ting Cao Microsoft Research
16:00
20m
Talk
Co-optimizing Memory-Level Parallelism and Cache-Level Parallelism
PLDI Research Papers
Xulong Tang Penn State, Mahmut Taylan Kandemir Pennsylvania State University, USA, Mustafa Karakoy TOBB University of Economics and Technology, Turkey, Meenakshi Arunachalam Intel, USA
Media Attached
16:20
20m
Talk
Low-Latency Graph Streaming using Compressed Purely-Functional Trees
PLDI Research Papers
Laxman Dhulipala Carnegie Mellon University, Guy E. Blelloch Carnegie Mellon University, Julian Shun MIT
16:40
20m
Talk
Composable, Sound Transformations of Nested Recursion and Loops
PLDI Research Papers
Kirshanthan Sundararajah Purdue University, Milind Kulkarni Purdue University
Media Attached