Observations and opportunities in architecting shared virtual memory for heterogeneous systems

@article{Vesel2016ObservationsAO,
  title={Observations and opportunities in architecting shared virtual memory for heterogeneous systems},
  author={J{\'a}n Vesel{\'y} and Arkaprava Basu and Mark Oskin and Gabriel H. Loh and Abhishek Bhattacharjee},
  journal={2016 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)},
  year={2016},
  pages={161-171}
}
Computing is becoming increasingly heterogeneous with accelerators like GPUs being tightly integrated with CPUs on the same die. Extending the CPU's virtual addressing mechanism to these accelerators is a key step in making accelerators easily programmable. In this work, we analyze, using real-system measurements, shared virtual memory across the CPU and an integrated GPU. We make several key observations and highlight consequent research opportunities: (1) servicing a TLB miss from the GPU can… CONTINUE READING

Citations

Publications citing this paper.
SHOWING 1-10 OF 47 CITATIONS

Improving and complementing virtual memory using hardware techniques

VIEW 12 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED

Analyzing memory management methods on integrated CPU-GPU systems

VIEW 17 EXCERPTS
CITES BACKGROUND
HIGHLY INFLUENCED

Interference from GPU System Service Requests

VIEW 6 EXCERPTS
CITES BACKGROUND & METHODS

Neighborhood-Aware Address Translation for Irregular GPU Applications

VIEW 6 EXCERPTS
CITES METHODS & BACKGROUND

Scalable and Efficient Virtual Memory Sharing in Heterogeneous SoCs with TLB Prefetching and MMU-Aware DMA Engine

VIEW 20 EXCERPTS
CITES BACKGROUND
HIGHLY INFLUENCED

Scheduling Page Table Walks for Irregular GPU Applications

VIEW 9 EXCERPTS
CITES BACKGROUND & METHODS

Reducing address translation overheads with virtual caching

VIEW 5 EXCERPTS
CITES BACKGROUND, RESULTS & METHODS
HIGHLY INFLUENCED

Near-Memory Address Translation

VIEW 8 EXCERPTS
CITES BACKGROUND, METHODS & RESULTS
HIGHLY INFLUENCED

Reducing GPU Address Translation Overhead with Virtual Caching

VIEW 23 EXCERPTS
CITES BACKGROUND, METHODS & RESULTS
HIGHLY INFLUENCED

References

Publications referenced by this paper.
SHOWING 1-10 OF 23 REFERENCES

CAPI: A Coherent Accelerator Processor Interface

VIEW 1 EXCERPT

Large pages and lightweight memory management in virtualized environments: Can you have it both ways?

VIEW 1 EXCERPT

Prediction-based superpage-friendly TLB designs

VIEW 1 EXCERPT

Redundant Memory Mappings for fast access to large memories

VIEW 2 EXCERPTS

Supporting superpages in non-contiguous physical memory

VIEW 1 EXCERPT

Unlocking bandwidth for GPUs in CC-NUMA systems

VIEW 1 EXCERPT

Efficient Memory Virtualization: Reducing Dimensionality of Nested Page Walks

VIEW 1 EXCERPT

Increasing TLB reach by exploiting clustering in page translations

VIEW 2 EXCERPTS