cfaed Seminar Series

Tal Ben-Nun , The Hebrew University of Jerusalem

Memory Access Patterns: The Missing Piece of the GPU Programming Puzzle

26.04.2016 (Tuesday) , 11:00 - 13:00
Andreas-Pfitzmann-Bau, Raum 1004 , Nöthnitzer Str. 46 , 01187 Dresden

GPUs play an increasingly important role in high-performance computing, accelerating
applications ranging from machine learning and computer vision to quantitative finance. In spite
of their popularity, GPU programming remains a challenging task. Manual management of the
multi-level memory hierarchy, inter-device synchronization, and various development
constraints often generate lengthy, error-prone, and architecture-specific code. Therefore, it is
imperative to develop new programming paradigms for efficient utilization of GPU and multi-
GPU systems, without sacrificing code simplicity and portability.

The talk will present Memory-Oriented Programming (MOP) - a data-centric programming
model organized around memory access patterns. MOP categorizes parallel algorithms by their
input and output parameters, where each parameter is characterized by a single pattern. Using
these patterns, which cover most existing GPU applications, we show that it is possible to
write short and intelligible code that attains high performance on a variety of GPU architectures
and multi-GPU nodes.
Furthermore, we demonstrate that the model can be used to increase the resilience of multi-
GPU nodes by providing transparent failure protection mechanisms.
We show that the resulting code is on par, and in some cases outperforms, existing manually
optimized production-grade libraries, exhibiting near-linear speedup on a single node with
multiple GPUs.
Performance is measured on fundamental computational operations, as well as real-world
applications in deep learning and non-negative matrix factorization.

Everybody is welcome!

Go back