
Seminar in Comp. Arch. - S1: ColumnDisturb & ABACuS (Spring 2026)
The seminar introduced column disturbance, a newly documented read‑disturbance effect in modern DRAM. Presenter Yong Jo explained the open‑bitline architecture of DDR4 and HBM2 chips, then described how aggressive activation of a single row can perturb bit‑lines across an entire subarray and even neighboring subarrays, causing bit flips far beyond the localized impact of classic row‑hammer or row‑press attacks. Experimental results showed the phenomenon is ubiquitous: all 216 DDR4 and four HBM2 chips tested exhibited column‑disturbance flips, and the time to first error shrank dramatically in newer die revisions—up to five‑fold faster in some 8 GB SK Hynix parts. In the worst case, a flip occurred in 63.6 ms, well within DDR4’s 64 ms refresh interval, meaning the error can happen before the next mandatory refresh. The data also revealed that column disturbance generates far more bit flips than row‑hammer, row‑press, or ordinary retention failures—up to 35× higher in certain Samsung devices. Flips were concentrated in the aggressor subarray’s shared bit‑lines, with neighboring subarrays experiencing roughly half the rate due to the open‑bitline sharing. Moreover, the direction of flips (1→0 versus 0→1) differed from other disturbances, underscoring a distinct physical mechanism. These findings imply that existing DRAM reliability models and mitigation techniques (e.g., targeted refresh or row‑hammer detection) are insufficient. Designers may need to revisit open‑bitline layouts, introduce more aggressive refresh schedules, or develop hardware‑level monitoring to detect column‑wide activity. For data‑center operators and system architects, the risk of silent data corruption grows as process nodes shrink, making proactive countermeasures essential.

"Can We Do Better?" Prof. Onur Mutlu's MICRO 2025 Keynote Talk at Seoul - 21.10.2025
Prof. Onur Mutlu’s MICRO 2025 keynote, titled “Can We Do Better?”, framed the memory bottleneck as the central obstacle to energy‑efficient, high‑performance computing. He argued that while the industry has long treated computing as an energy problem, the majority of...

Digital Design & Comp. Arch: L15: Dataflow, Superscalar Execution & Branch Prediction (Spring 2026)
On 17 April 2026, Prof. Onur Mutlu delivered ETH Zürich’s Lecture 15 on dataflow, superscalar execution, and branch prediction as part of the Digital Design & Computer Architecture spring series. The session unpacked how dataflow architectures expose fine‑grained parallelism, how superscalar CPUs...

AVATAR: A Variable-Retention-Time Aware Refresh for DRAM Systems - DSN 2025 Test-of-Time Award
The DSN 2025 Test‑of‑Time award honored the seminal AVATAR paper, which tackled the growing DRAM refresh burden as capacities and operating frequencies increased. The authors highlighted that traditional uniform refresh intervals ignore the non‑uniform, variable‑retention‑time (VRT) behavior of memory cells,...

Memory-Centric Computing: Enabling Fundamentally-Efficient Computers - Georgia Tech ECE Seminar
Professor Honor Mutlu’s Georgia Tech seminar highlighted a fundamental shift in computer architecture: moving from processor‑centric designs to memory‑centric computing. He argued that exploding data volumes in AI, genomics, and scientific domains have turned data movement into the primary performance...

Seminar in Comp. Arch. - L7: Virtual Memory (Spring 2026)
On 2 April 2026, ETH Zürich’s Computer Architecture seminar hosted a deep dive into virtual memory, led by Konstantinos Kanellopoulos and Prof. Onur Mutlu. The session combined foundational concepts with cutting‑edge research on processing‑in‑memory, memory‑centric architectures, and security vulnerabilities such as RowHammer. Attendees received a rich...

Digital Design & Comp. Arch: L13: Precise Exceptions & Interrupts (Spring 2026)
The Spring 2026 lecture on Digital Design & Computer Architecture focuses on precise exceptions and interrupts, explaining how they preserve sequential semantics in pipelined processors and set the stage for out‑of‑order execution. The instructor shows that modern pipelines contain multiple functional...

Memory-Centric Computing: Recent Advances in Processing-in-DRAM: IEDM Invited Talk - 09.12.2024
The invited IEDM talk highlighted memory‑centric computing as a response to exploding data volumes and the growing energy cost of moving that data. The speaker argued that today’s processor‑centric designs waste up to 90% of system energy on memory accesses,...

Understanding & Designing Modern Storage Systems - M5: Processing Inside NAND Flash Memory
The video introduces FlashCosmos, a new in‑flash processing technique that performs bulk bitwise operations directly inside NAND flash memory. Presented as part of a recent MICRO 2022 paper, the work targets the growing data‑movement bottleneck that hampers databases, graph analytics,...

MCCSys-5: 5th Workshop on Memory-Centric Computing Systems, Held with ASPLOS 2026 - 23 March 2026
The fifth Memory‑Centric Computing Systems (MCCSys‑5) workshop, co‑located with ASPLOS 2026, gathered researchers to confront the growing memory bottleneck that now dominates performance, energy consumption, and hardware cost across data‑intensive workloads. Organizers outlined the agenda—keynotes on memory‑centric architectures, recent advances...

ASPLOS 2026: 2nd Workshop on Virtuoso: Ideas and Infrastructures for Novel HW/OS Interfaces
The 2nd Virtuoso workshop, part of ASPLOS 2026, will convene on March 23, 2026 in Pittsburgh to explore hardware‑software co‑design for memory management. Organized by CMU‑SAFARI, the full‑day event includes a tutorial, presentations of recent research, and hands‑on sessions using...

Seminar in Comp. Arch. - L6: Machine Learning-Driven Memory and Storage System Design (Spring 2026)
On March 26, 2026, ETH Zürich’s Computer Architecture seminar featured a deep dive into machine‑learning‑driven memory and storage system design, presented by Rahul Bera, Rakesh Nadig, and Prof. Onur Mutlu. The session highlighted how AI techniques can automate tiered memory management,...

Revisiting RowHammer - Top Picks in Hardware and Embedded Security - Prof. Onur Mutlu - 30.10.2025
The talk revisits the seminal Rowhammer problem, presenting the 2020 "Revisiting Rowhammer" paper that conducted the largest experimental study to date on real DRAM chips. By testing roughly 1,600 devices from three major vendors across DDR3, DDR4, and LPDDR4 generations,...

Digital Design & Comp. Arch: L10: Microarchitecture Fundamentals and Design (Spring 2026)
The lecture introduces micro‑architecture fundamentals by contrasting single‑cycle and multicycle processor designs. It explains that a single‑cycle processor implements the entire ISA instruction path—fetch, decode, execute, memory access, and write‑back—in one combinational pass, requiring the clock period to accommodate the...

Memory System Design for AI/ML & ML/AI for Memory System Design - SRC AIHW Annual Review - 23.07.24
The SRC AIHW annual review highlighted a critical challenge in modern AI/ML systems: data movement consumes the majority of system energy, especially in large‑scale models running on edge TPUs where over 90% of power is spent on off‑chip interconnects. The...