Interested in joining our group? We are currently recruiting motivated graduate students!
Overview
The Systems Optimization and Computer Architecture Lab (SoCal) focuses on a broad range of computer systems-related research, spanning data centers to microarchitecture.
With the emergence of cloud-scale computing and the impeding end of Moore's law, computer systems innovation faces unprecedented obstacles across the entire hardware-software stack.
With the slowdown in technology scaling, the onus of future power and performance improvements falls solely on microarchitecture and software innovations to make more efficient use of the hardware.
Similarly, at the data center scale, novel innovations are required to make efficient use of commodity hardware in the face of data-intensive workloads, heterogenous hardware, and energy-constrained infrastructures.
Below is a list of current research projects in the SoCal lab:
Projects
Energy Efficient Data Centers
- Coordinating network and server energy savings [IPDPS'18]
- QoS-aware dynamic power management [ISLPED'16]
- Reliability-aware scheduling [DAC'16]
- Energy proportional computing [MICRO'12, HPCA'14, IISWC'15, ISCA'16]
GPU Microarchitecture
- Data-dependent parallelism support [MICRO'17]
- Approximate computing support [HPCA'16]
- Efficient power gating of execution units [MICRO'13, ICS'16]
Systems Support for Heterogeneous Accelerators
- Accelerator management runtimes
- Multi-GPU programming support
- GPU accelerated NFV