13436

Unlocking Bandwidth for GPUs in CC-NUMA Systems

Neha Agarwal, David Nellans, Mike O’Connor, Stephen W. Keckler, Thomas F. Wenisch
University of Michigan
2015 International Symposium on High Performance Computer Architecture (HPCA), 2015

@article{agarwal2015unlocking,

   title={Unlocking Bandwidth for GPUs in CC-NUMA Systems},

   author={Agarwal, Neha and Nellans, David and O’Connor, Mike and Keckler, Stephen W and Wenisch, Thomas F},

   year={2015}

}

Download Download (PDF)   View View   Source Source   

1160

views

Historically, GPU-based HPC applications have had a substantial memory bandwidth advantage over CPU-based workloads due to using GDDR rather than DDR memory. However, past GPUs required a restricted programming model where application data was allocated up front and explicitly copied into GPU memory before launching a GPU kernel by the programmer. Recently, GPUs have eased this requirement and now can employ on-demand software page migration between CPU and GPU memory to obviate explicit copying. In the near future, CCNUMA GPU-CPU systems will appear where software page migration is an optional choice and hardware cache-coherence can also support the GPU accessing CPU memory directly. In this work, we describe the trade-offs and considerations in relying on hardware cache-coherence mechanisms versus using software page migration to optimize the performance of memory-intensive GPU workloads. We show that page migration decisions based on page access frequency alone are a poor solution and that a broader solution using virtual address-based program locality to enable aggressive memory prefetching combined with bandwidth balancing is required to maximize performance. We present a software runtime system requiring minimal hardware support that, on average, outperforms CC-NUMA-based accesses by 1.95x, performs 6% better than the legacy CPU to GPU memcpy regime by intelligently using both CPU and GPU memory bandwidth, and comes within 28% of oracular page placement, all while maintaining the relaxed memory semantics of modern GPUs.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2017 hgpu.org

All rights belong to the respective authors

Contact us: