Title: Performance is All About Instruction Supply
Abstract: The computing landscape is shifting from the traditional desktop computing to mobile-cloud computing model. In this new model, datacenters have emerged as the workhorses that do all the heavy-duty computations and serve as the backbone of mobile services. However, the processors deployed in datacenters remain highly underutilized due to a large mismatch in application characteristics and processor microarchitecture. In this talk I will make a case for specializing server processors to application characteristics. Concretely, datacenter applications exhibit massive instruction footprints that frequently stall the core front-end. I will discuss microarchitectural mechanisms to mitigate this front-end bottleneck. I will also touch upon our recent work on energy-efficient instruction scheduling and uncovering hidden instructions that can be a huge security loophole.
Speaker Bio: Rakesh Kumar is an Associate Professor at NTNU, Norway. His work focuses on scale-out servers, hardware/software co-designed processors, memory system, and run-time code generation and optimizations. He received his PhD from UPC Barcelona in 2014. During his internship at Intel Barcelona Research Center he developed memory controllers for Intel Skylake server architecture.