HUM Hidden Unified Memory

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).

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.

HUM can be downloaded from GitHub.

• https://github.com/mcrl/hum

If you would like to download the HUM user manual, please email us.

Jaehoon Jung, Jungho Park, and Jaejin Lee

hum@aces.snu.ac.kr

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
      https://thunder.snu.ac.kr
    
    Contributors:
      Jaehoon Jung, Jungho Park, and Jaejin Lee