General Purpose Task-Dependence Management Hardware for Task-Based Dataflow Programming Models

Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at runtime. Managing these dependences introduces noticeable overheads when targeting fine-grained tasks, diminishing the potential speedups or even i...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings - IEEE International Parallel and Distributed Processing Symposium pp. 244 - 253
Main Authors: Xubin Tan, Bosch, Jaume, Vidal, Miquel, Alvarez, Carlos, Jimenez-Gonzalez, Daniel, Ayguade, Eduard, Valero, Mateo
Format: Conference Proceeding
Language:English
Published: IEEE 01.05.2017
Subjects:
Discrete cosine transforms
Energy saving
Fine- grained task parallelism
Hardware
Instruction sets
Nested tasks
Parallel processing
Picos++ Hardware Accelerator
Programming
Runtime
Task-based dataflow programming models
ISSN:1530-2075
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Task-based programming models such as OpenMP, IntelTBB and OmpSs offer the possibility of expressing dependences among tasks to drive their execution at runtime. Managing these dependences introduces noticeable overheads when targeting fine-grained tasks, diminishing the potential speedups or even introducing performance losses. To overcome this drawback, we present a general purpose hardware accelerator, Picos++, to manage the inter-task dependences efficiently in both time and energy. Our design also includes a novel nested task support. To this end, a new hardware/software co-design is presented to overcome the fact that nested tasks with dependences could result in system deadlocks due to the limited amount of resources in hardware task dependence managers. In this paper we describe a detailed implementation of this design and evaluate a parallel task-based programming model using Picos++ in a Linux embedded system with two ARM Cortex-A9 and a FPGA. The scalability and energy consumption of the real system implemented have been studied and compared against a software runtime. Even in a system limited to 2 threads, using Picos++ results in more than 1.8x speedup and 40% of energy savings in the most demanding parallelizations of real benchmarks. As a matter of fact, a hardware task dependence manager should be able to achieve much higher speedup and provide more energy savings with more threads.
ISSN:1530-2075
DOI:10.1109/IPDPS.2017.48