CMU-SAFARI / pLUTo

pLUTo is a DRAM-based Processing-using-Memory architecture that leverages the high density of DRAM to enable the massively parallel storing and querying of lookup tables (LUTs)

Home Page:https://arxiv.org/abs/2104.07699

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

pluto-logo

Enabling Massively Parallel Computation in DRAM via Lookup Tables

GitHub GitHub Release
MICRO DOI arXiv DOI Zenodo DOI

Table of Contents
  1. What is pLUTo?
  2. Citation
  3. Repository Structure
  4. Execution Instructions
  5. License
  6. Contact
  7. Acknowledgments

What is pLUTo?

pLUTo is a Processing-using-Memory architecture that enables the execution of complex operations in-memory. To achieve this, the key idea of pLUTo is to replace complex operations with low-cost, massively parallel table lookups that produce the same result.

The pLUTo LUT Query operation is at the core of pLUTo's functionality: this operation enables every element in an input memory row to be used to query a LUT that may contain up to as many entries as the number of rows in a DRAM subarray.

Citation

Please cite our full MICRO 2022 paper if you find this repository useful.

João Dinis Ferreira, Gabriel Falcao, Juan Gómez-Luna, Mohammed Alser, Lois Orosa, Mohammad Sadrosadati, Jeremie S. Kim, Geraldo F. Oliveira, Taha Shahroodi, Anant Nori, and Onur Mutlu. "pLUTo: Enabling Massively Parallel Computation in DRAM via Lookup Tables". In Proceedings of the 55th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO), 2022. https://doi.org/10.1109/MICRO56248.2022.00067.

You can also use the following BibTeX entry for this purpose:

@inproceedings{ferreira2022pluto,
  author = {Ferreira, João Dinis and Falcao, Gabriel and Gómez-Luna, Juan and Alser, Mohammed and Orosa, Lois and Sadrosadati, Mohammad and Kim, Jeremie S. and Oliveira, Geraldo F. and Shahroodi, Taha and Nori, Anant and Mutlu, Onur},
  title = {{pLUTo: Enabling Massively Parallel Computation in DRAM via Lookup Tables}},
  doi = {10.1109/MICRO56248.2022.00067},
  booktitle = {Proceedings of the 55th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO)},
  year = {2022}
}

Repository Structure

This repository contains the artifact evaluation materials associated with pLUTo: Enabling Massively Parallel Computation In DRAM via Lookup Tables, published at MICRO 2022.

We provide source code and helper files to reproduce:

  1. SPICE simulations (shown in Figure 6 in the paper)
  2. The simulation of pLUTo's operation (shown in Figures 7-13 in the paper)

We provide in-depth instructions on how to reproduce each of the artifacts in the README.md files in the pluto_sim and spice directories.

The following is a high-level overview of the structure of this repository. We highlight select folders and files of note.

.
+-- LICENSE
+-- README.md
+-- pluto_sim/
|   +-- baselines                   -->             baseline results for the evaluated workloads
|   +-- pysim                       -->             output folder (empty when the repository is first cloned)
|   +-- pysim_reference             -->             reference results for the pLUTo-based execution of the evaluated workloads
|   +-- LUT_Loading_Times.xlsx      -->             used to produce the results shown in Figure 11 in the paper
|   +-- README.md                   --> START HERE: step-by-step instructions on how to reproduce our results
|   +-- requirements.txt            -->             Python requirements to be installed prior to execution
|   +-- sim_walkthrough.ipynb       -->             interactive Python Jupyter Notebook with step-by-step instructions on how to reproduce our results
|   +-- ...
+-- resources/
|   +-- ...                         -->             media rendered in this README
+-- spice/
|   +-- media/                      -->             media rendered in the README
|   +-- out/                        -->             output folder
|   +-- runs/                       -->             output folder
|   +-- README.md                   --> START HERE: step-by-step instructions on how to reproduce our results
|   +-- pluto-a.asc                 -->             cell design file
|   +-- pluto-b.asc                 -->             cell design file
|   +-- pluto-c.asc                 -->             cell design file
|   +-- reference.asc               -->             cell design file
|   +-- transistor_model.pm         -->             transistor model

Execution Instructions

For step-by-step instructions on how to reproduce our results, please see the README in the spice folder and the README in the pluto_sim folder.

License

Distributed under the MIT License. See LICENSE for more information.

Contact

João Dinis Ferreira - hello@joaof.eu

Acknowledgements

We acknowledge support from SAFARI Research Group's industrial partners.

About

pLUTo is a DRAM-based Processing-using-Memory architecture that leverages the high density of DRAM to enable the massively parallel storing and querying of lookup tables (LUTs)

https://arxiv.org/abs/2104.07699

License:MIT License


Languages

Language:Jupyter Notebook 69.4%Language:AGS Script 14.9%Language:Python 10.8%Language:Raku 4.9%