HUM Hidden Unified Memory

Overview

HUM (Hidden Unified Memory) is an open-source framework for NVIDIA GPUs that hides host-to-device memory copy time without any code modification. It overlaps the host-to-device memory copy with host computation or CUDA kernel computation by exploiting CUDA Unified Memory and fault mechanisms.

HUM consists of two components: a linux kernel module (HUM-driver), and a runtime system (HUM-runtime).

Publications

Please use the citation below to reference HUM.

  • [PPoPP] Jaehoon Jung, Daeyoung Park, Youngdong Do, Jungho Park, and Jaejin Lee. Overlapping Host-to-Device Copy and Computation using Hidden Unified Memory, PPoPP '20: Proceedings of the 25th Symposium on Principles and Practice of Parallel Programming, pp. 321—335, San Diego, California, USA, February 2020.

Download

HUM can be downloaded from GitHub.

Manual

If you would like to download the HUM user manual, please fill out the following form and click the download button. An email containing the download URL will be sent to the email address.

Contributors

  • Current: Jaehoon Jung, Jungho Park, and Jaejin Lee

Contact and Bug Report

E-mail: hum@aces.snu.ac.kr

License

Copyright (c) 2020 Seoul National University.                            
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
	1. Redistributions of source code must retain the above copyright
		 notice, this list of conditions and the following disclaimer.
	2. Redistributions in binary form must reproduce the above copyright
		 notice, this list of conditions and the following disclaimer in the
		 documentation and/or other materials provided with the distribution.
	3. Neither the name of Seoul National University nor the names of its
		 contributors may be used to endorse or promote products derived
		 from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY SEOUL NATIONAL UNIVERSITY "AS IS" AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL SEOUL NATIONAL UNIVERSITY BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.

Contact information:
	Center for Manycore Programming
	Department of Computer Science and Engineering
	Seoul National University, Seoul 08826, Korea
	http://aces.snu.ac.kr

Contributors:
	Jaehoon Jung, Jungho Park, and Jaejin Lee