News

Subscribe (RSS)

2026

  1. Oct 30

    Welcome, James Chuang and Madhur Kumar

    Two new Ph.D. students joined the STAR Lab in 2026.

  2. Oct 30

    Junsu Kim wins the J.K. Zee Fellowship

    Ph.D. student Junsu Kim has received the J.K. Zee Fellowship at UBC.

  3. Oct 29

    ASPLOS 2027 acceptance: MagicFlow

    Our paper on the magic state bottleneck in fault-tolerant quantum computing has been accepted to ASPLOS 2027.

  4. Oct 28

    Keynote at the IBM AI Compute Symposium

    Keynote on 3D-DRAM-based accelerators for generative inference at the IBM IEEE CAS/EDS AI Compute Symposium.

  5. Jul 10

    Keynote at SACMAT 2026

    Keynote on resilient DRAM for secure and intelligent computing at the ACM Symposium on Access Control Models and Technologies.

  6. Jul 9

    On Techquickie: Why PC Memory is Stuck in 1995

    Interviewed by Techquickie (Linus Media Group) on why PCs keep CPU and GPU memory separate, and what unified memory would change.

  7. Jun 30

    Service: General Co-Chair of ISCA 2027

    Serving as General Co-Chair of the International Symposium on Computer Architecture (ISCA 2027).

  8. Jun 25

    ISCA 2026 Distinguished Artifact Award

    Meng Wang's ISCA 2026 artifact was one of two Distinguished Artifacts.

  9. Jun 20

    ISCA 2026: Three papers, including the first silicon results for Raptor

    Three papers at ISCA 2026 spanning 3D-DRAM accelerators, fault-tolerant quantum computing, and RowHammer defense.

  10. Apr 30

    Recorded talk at ETH Zürich: Scaling the Memory Wall

    A SAFARI Live Seminar at ETH Zürich on 3D-DRAM-based accelerators for generative inference, now available as a recording.

  11. Apr 15

    NSERC Discovery Grant: Memory-Centric Architectures for Large-Scale ML Inference

    Awarded a five-year NSERC Discovery Grant (2026–2031).

2025

  1. Sep 18

    NeurIPS 2025 acceptance: Foresight (adaptive layer reuse for text-to-video generation)

    Our work on adaptive layer reuse for fast, high-quality text-to-video generation has been accepted to NeurIPS 2025.

  2. Jul 14

    MICRO 2025 acceptance: RayN (ray tracing acceleration with near-memory computing)

    Our MICRO 2025 paper introduces RayN: near-memory computing for ray tracing acceleration.

  3. Jun 28

    DSN 2025 Test-of-Time Award and ISCA Hall of Fame

    Our DSN paper received the DSN 2025 Test-of-Time Award, and I was inducted into the ISCA Hall of Fame.

  4. Jun 21

    CBP2025 @ ISCA 2025: Bullseye (a hybrid branch predictor)

    Our team participated in CBP2025 with Bullseye, a hybrid branch predictor.

  5. Jun 21
  6. May 1

    ICML 2025 acceptance: Dialogue Without Limits (constant-sized KV caches for long responses)

    ICML 2025 accepted our work on constant-sized KV caches to enable extended LLM responses.

  7. Apr 9

    New arXiv preprint: Context Switching for Secure Multi-programming of Near-Term Quantum Computers

    We posted a new arXiv preprint on context switching for secure multi-programming of near-term quantum computers.

  8. Mar 22

    ISCA 2025: Four accepted papers across RowHammer, quantum simulation, prefetching, and sparse compute

    ISCA 2025 accepted four of our papers spanning RowHammer, noisy quantum simulation, prefetching, and sparse accelerators.

  9. Mar 5

    HPCA 2025: Two RowHammer-defense papers (including a Distinguished Artifact Award)

    At HPCA 2025, we presented two papers on RowHammer defense; QPRAC received a Distinguished Artifact Award.

  10. Feb 10

    Service: External Review Committee member for MICRO 2025

    Serving on the MICRO 2025 external review committee.

  11. Feb 3

    Service: Program Committee member for HPCA 2025 and ISCA 2025

    Serving on the HPCA 2025 and ISCA 2025 program committees.

  12. Jan 15

    Service: Associate Editor, IEEE Computer Architecture Letters

    Joined the editorial board of IEEE Computer Architecture Letters as an Associate Editor.

  13. Jan 10

    DATE 2025 acceptance: Slipstream (semantic training acceleration for recommendation models)

    DATE 2025 accepted our paper on semantic-based training acceleration for recommendation models.