Linpeng Tang will present his FPO, "Efficient Processing and Delivery of Multimedia Data" on Friday, 2/23/2018 at 1:30pm in CS 402. The members of this committee is as follows: Kai Li (Adviser); Examiners: Jennifer Rexford, Kyle Jamieson, and Kai Li; Readers: Wyatt Lloyd and Michael Freedman Everyone is invited to attend. His abstract follows below. The explosion of multimedia data on the Internet in recent years has greatly enriched people's online experience. However, it also poses a great challenge to analyze and process, and then deliver such content to the worldwide audience. This dissertation presents novel approaches to improve the overall efficiency of the stack by novel system designs. First, to improve the caching performance of the flash-memory caches for content delivery network, this thesis proposes RIPQ, a framework for efficient and advanced caching with flash memory. Traditional implementations of these algorithms generate random writes that are not appropriate for flash devices, decreasing the device's performance and lifespan. RIPQ overcomes this issue by aggregating small writes, colocating items with similar priorities, and perform lazy updates to achieve low overhead. Second, this thesis proposes Chess, which uses popularity prediction for higher quality video streaming. Although better encodings improve video streaming, they are also compute-intensive, and it's infeasible to encode all videos uploaded to Facebook with the highest quality codec. However, because the accesses to videos are highly skewed, we may obtain most of the benefit by only running the compute- intensive encoding on a small portion of popular videos. Chess meets this demand by designing an approximate but fast base predictor with the access history information, and using an online learning method to combine multiple such predictors as well as the social signals to boost accuracy. Lastly, this thesis investigates how to accelerate deep learning models on manycore CPUs. Deep learning is now widely used for analyzing multimedia data, but it's compute-intensive, which constitutes its major bottleneck. The manycore CPU, combining both high FLOPS and a flexible computing model, is a promising solution to this problem. However, existing frameworks are still mainly optimized for GPU, and don't run efficiently on this architecture. To overcome this issue, this thesis proposes Graphi, the first attempt to accelerate the execution of computation graphs for deep learning models on this architecture. Graphi determines the optimal parallel settings with a profiling step, runs concurrent operations with low contention, and further reduces execution makespan with critical-path first scheduling.