Literature Survey

ECE 587, Spring 2025



Progress Report Due: 04/06 (Sun.), by the end of the day (Chicago time)
Final Survey Paper Due: 05/04 (Sun.), by the end of the day (Chicago time)
Late submissions will NOT be graded

Below I have listed a few articles related to the ideas of hardware/software co-design that we have explored in the class. You should choose one from them, read it, and write your survey paper based on it. Note that to survey the article you have chosen, you may need to read its references and references of references to understand the background and the previous works.

You could also choose your own article out of my list. However, if you prefer to do so, you should email me your choice no later than the end of 03/28 (Fri.) and I will decide if your choice is appropriate. Generally speaking, your choice should be related to hardware/software co-design, not used for ECE 587 literature survey from a previous semester, and published in leading IEEE/ACM conferences or journals within the last 10 years.

    (For IEEE papers, please click Institutional Sign In to access from outside of IIT network.)
  1. Lamport, The Parallel Execution of DO Loops, CACM 1974.
  2. Batson et al., High-Level Specifications: Lessons from Industry, LNCS 2852.
  3. Hoe et al., Operation-Centric Hardware Description and Synthesis, TCAD 2004.
  4. Chen et al., FCUDA-NoC: A Scalable and Efficient Network-on-Chip Implementation for the CUDA-to-FPGA Flow, TVLSI 2015.
  5. Izraelevitz et al., Reusability is FIRRTL Ground:Hardware Construction Languages, Compiler Frameworks, and Transformations, ICCAD 2017.
  6. Liu et al., Optimizing CNN Model Inference on CPUs, USENIX ATC 2019.
  7. Du et al., Accelerating DNN Inference with Heterogeneous Multi-DPU Engines, DAC 2023
  8. Jiang et al., BlueFace: Integrating an Accelerator into the Core's Pipeline through Algorithm-Interface Co-Design for Real-Time SoCs, DAC 2023
  9. Hu et al., Fast and Scalable Gate-Level Simulation in Massively Parallel Systems , ICCAD 2023.
  10. Ghannane et al., DiviML: A Module-based Heuristic for Mapping Neural Networks onto Heterogeneous Platforms, ICCAD 2023.

The purpose of the literature survey is to explain the research work in the chosen article. Therefore, I would recommend the following sections to be included in your survey paper.

  1. Introduction: discuss briefly what the research is generally about. Compare briefly this article to previous works.
  2. Background: present the common assumptions used in the article and its references based on mathematical models.
  3. Problem Formulation (and Previous Works): show the exact problem this article attempts to solve. If there are previous works attempting to solve the similar problems, make it clear what are the differences.
  4. Proposed Approach and Experimental Results: provide details on the approach the article taken to solve the problem. Summarize the experimental results.
  5. Conclusion and Future Works: present the conclusion and discuss the problems you are interested to investigate in the future after reading the article. You don't need to have a solution yet.
  6. References: list all the articles you have read during the survey process. All your writings should be accompanied by proper citations to the references here.

Overall, your survey paper should "look like" the research articles you have surveyed. You should make use of figures and tables to present ideas and results better. It is up to you to use a single or double column format but your survey paper should NOT EXCEED 6 pages. Use common sense to choose proper fonts, line spacing, and paper margins.

Your survey paper contributes 30 points to your final grade.

A progress report (10 points) is due 04/06, indicating that you have chosen the article you want to survey and have made progresses collecting necessary background information. The progress report should include Background, Problem Formulation, and References sections. It will be evaluated mostly based on completion.

The final survey paper (20 points) is due 05/04. It will be evaluated solely based on the quality of your writing as follows:

We will use Turnitin to monitor your submissions on Canvas. To avoid possible false positives on references, for both the progress report and the final survey paper, please separate your submission into two files, one for the main content, and one for the references. You will receive 0 point if Turnitin shows a similarity score of at least 50% for your main content file, or if you use any mechanism to bypass Turnitin, for example, to convert texts into images. You may receive up to 25% penalty if Turnitin shows a similarity score between 25% and 50%. There will be no penalty if Turnitin shows a similarity score no more than 25%. Please rephrase sentences as needed and avoid direct quotations of sentences or paragraphs.

The survey should be done individually. You can discuss the article with other students but all the writings should be your OWN. COPY without proper CITATION, and extensive COPY from other materials including but not limited to the article your surveyed, the references, and the textbook will be treated as PLAGIARISM and called for DISCIPLINARY ACTION. NEVER share your reports with others.