I have been a PhD student in CSE@CUHK since 2022, advised by Prof. Ming-Chang Yang.

Interests. I am interested in processing-in-memory, and work on stuff related to it.

Recent Preprints. Here are some papers/preprints I was or am currrently working on.

  • "Extrating Core-Value Sets from Social Media: A Case Study on Meta"Title Only.
    Magic for getting some INFORMATION!

  • "Pragmatic Methods for Breaking Any Public-Key Encryption"Title Only.
    Magic for getting in ANYWHERE!

  • "Curved Layout, Asynchronous Stacking, and Elective Interconnects"Title Only.
    Magic for manipulating ELECTRONS!

  • "NVIDIABleed: When GPUs Miss Control-Flow Integrity? "Title Only.
    Magic for Control-Flow Hijacking, yet Reverse Engineering!
    In case you do not know: Data Hijacking and Corruption also!

  • My Comments. To make sure you know my demands, the incident is still not well-resolved. Credited to those two quantitative approaches, the culture is super toxic (example). Clearly, there are still a lot of politicians, who obtain funding and never do their jobs. And I already run out of my patience. Why don't we just expect a public statement on the issue (and ideally, the list who shall be permenantly out) from SIGARCH? And, before the community drops their vanity, all funding to them shall be cut, all job offers to their students shall be put on hold, and all relevant events shall not be supported. Before seeing the statement, I would keep throwing out new stuff. Clearly, if none of you care, I won't care as well.

    Paper-accepting crew at HPCA 2025. It seems that HPCA 2025 does not allow the authors to identify reviewers (though it is too clear for me after reading reviews, since I do read others' papers). So I name those academic politicians hereby.
    Group 1 Head, followed by his students (like this and this);
    Group 2 Head, followed by his collegues (like this, this and this);
    and the incremental unit, the yet another quantitative approach and the ASSPLOSer.
    All center around the storager plus his buddy; and the chair professor.
    Funny. Please enjoy accepting each other's papers! More interesting stuff would come soon!

  • "Functionality Locality, Mixture & Control = Logic = Memory"Revising.

  • "Value, Representation, Information and Communication"

  • "Reversed Indexes ≈ Values in Wavelet Trees"

    Finalized Papers. Some papers are rewritten (and implications are added).

  • "Pareto Improvements for Any Services: The Succincter Cluster"Abstract Only.

  • "Rethinking Instruction Set Architecture and Compiler for Bit-serial SIMD Processing Using Memory in DRAM"Abstract Only.
    Clearly, something else happens; and more obviously, someone is wrong.

  • "Log-Polar Coordinates Made Efficient, Simple and Explainable"Abstract Only.
    Related preprints can be found here; and here.

  • "Feint Behaviors and Strategies: Formalization, Implementation and Evaluation"
    Related preprints can be found here for action formalization; and here for strategy design.

  • "SLITS: Sparsity-Lightened Intelligent Thread Scheduling"
    The conference version at SIGMETRICS 2023 is here, and the extended abstract is here.

  • "CHOPPER: A Compiler Infrastructure for Programmable Bit-serial SIMD Processing Using Memory in DRAM"
    The conference version at HPCA 2023 is here, and the summary is here.

    The full list of my published papers can be accessed hereby or hereby.

    Discussion. Since I am always happy to discuss about stuff or contribute to something useful, I prepare this so that we can have a discussion though we may not know each other in person.

    Disclaimer. I have abandoned this domain name (and this) since 2021, so I really do not have any ideas why it is automatically redirected: though I did find the contents interesting.

    "The disparity between problem solving and solving problems is the problem formalization." Therefore, this is precisely the reason why such an approach is much more effective.