Architecture of Computing Systems – ARCS 2020 by Unknown

Architecture of Computing Systems – ARCS 2020 by Unknown

Author:Unknown
Language: eng
Format: epub
ISBN: 9783030527945
Publisher: Springer International Publishing


6 Conclusion

We presented X-CEL, a measurement-based method to estimate the potential of near-memory acceleration. It helps to perform an early yet robust estimation whether the development effort of a near-memory accelerator is worthwhile. The two-stage method is based on measurements of an easy-to-integrate near-memory core (near-memory, but no accelerator) variant, which is closer to the target design than the existing baseline implementation (neither near-memory, nor hardware-accelerated). We showcased X-CEL with a (near-memory) graph copy problem in a tile-based MPSoC with a set of distributed graph algorithm kernels. An in-depth analysis revealed that the second stage estimate is within 5 % of the actual speedup in 70 % of the configurations. Moreover, it has 36 % higher accuracy than the original estimate.

Future work could refine the estimation model, as well as extend the framework to more case studies.

All in all, we envision X-CEL to become an x-cel-lent tool in the hand of developers to make sophisticated predictions on the near-memory acceleration potential and thereby avoid unnecessary development effort.

References

1.

Ahn, J., Hong, S., Yoo, S., Mutlu, O., Choi, K.: A scalable processing-in-memory accelerator for parallel graph processing. In: Proceedings of the 42nd Annual International Symposium on Computer Architecture, Portland, OR, USA, 13–17 June 2015, pp. 105–117 (2015). https://​doi.​org/​10.​1145/​2749469.​2750386



Download



Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.