Skip to content

Seminar: Holistic System Design — the Avenue Toward the Data-centric Future

February 12, 2021 - 12:00pm

EECS Seminar

Speaker: Hung-Wei Tseng

Title: Holistic System Design — the Avenue Toward the Data-centric Future

Zoom: https://ucmerced.zoom.us/j/86458710570?pwd=K1ZZWG5DOTJwenNrcFdFcG5FcVZFZz09

Abstract:

The rapid growth of datasets in applications significantly changes the landscape of modern computer architectures as they incorporate hardware accelerators, GPGPUs, non-volatile storage devices, in addition to traditional architectural components. However, these computers as well as programs running on top of them still inherit the half-century-old, layered system design that relies on the relatively slowly evolved CPU technologies. As a result, the overhead of layering has become an emerging bottleneck in performance.

To allow computers to make efficient use of modern heterogeneous computing units and data storage devices, and lead to performance improvements, we should revisit the entrenched layered system. Instead of local optimizations, we should take a holistic view of our system components when designing applications. To facilitate holistic system design but still maintain the benefits of layered design, we need to revisit the interface that each component exposes, remove redundant layers, and rethinking the roles of system components.

In this talk, Hung-Wei will demonstrate how holistic system design and the three philosophies can dramatically change the system performance. We can achieve a speedup of 2.5x for key-value store, accelerate approximate computing by 1.5x, and allow untethered VR gaming to be possible.
 

Bio:

Hung-Wei is currently an assistant professor in the Department of Electrical and Computer Engineering at the University of California, Riverside. He is now leading the Extreme Storage & Computer Architecture Laboratory and focusing on tackling the performance issues in modern heterogeneous computer systems through intelligent data storage. He is recognized by facebook faculty research award and IEEE Micro "Top Picks from Computer Architecture" in 2020 for his research in accelerating data-intensive applications through revisiting the storage system design. He got his Ph.D. from the Department of Computer Science and Engineering at the University of California, San Diego.