3D die-stacked memory offers the promise of increased memory bandwidth for today’s ever-scaling workloads. However, capacity and cost constraints often prevent building memory systems exclusively with stacked memory. As such, memory systems of future computers are likely to have a heterogeneous organization consisting of high-bandwidth 3D die-stacked memory and high-capacity memories. For this project, we explore different mechanisms to expose the bandwidth-benefits of stacked memory, while maintaining the capacity and cost-benefits of standard memories.
Datacenters today require a strong level of security and availability. Such systems need resilient memories, which are protected from both adversarial attacks and naturally occurring errors. Unfortunately, providing such guarantees involves an inherent cost in terms of performance and power, which often make these systems impractical. Our goal as computer architects, is to design scalable resilient memory systems that provide strong reliability and security, with minimal impact on performance or energy.
Quantum computing promises exponential speedups for a class of important problems. However, this potential can only be realized using large-scale quantum systems with a large number of qubits. Unfortunately, building a scalable quantum computer has several challenges that must be overcome, including the design of conventional computing and memory systems that can efficiently interface with the quantum substrate while obeying the thermal and power constraints dictated by the quantum devices. As computer architects, we try to address the system design challenges for scalable quantum computers.
To combat the data explosion, enterprise systems are seeing an increasing focus on storage and backup efficiency to achieve greater agility. SSD (Solid State Disk) storage serve as an ideal disk replacement or supplement, as they can achieve superior performance with lower standby power. However, the adoption of SSDs in these systems is hindered by the challenges of endurance, reliability, and performance predictability. We, as system architects, seek to address these limitations of SSD for datacenters and enterprise systems by exploring innovations spanning multiple levels of the compute stack.