Obvious ways to get performance wins are obvious. By the end of 2017, every .NET developer should know about concurrency and parallelism, ORMs and web response caching, efficient collections and garbage collection internals. But there are so many performance wins ripe for the picking that most .NET developers don’t know about.
This talk is all about squeezing the best performance out of modern hardware. And modern hardware is about vectorization in addition to parallelization; modern hardware is about optimizing your instruction flow and data layout for last-generation instruction sets and cache structures; modern hardware is understanding that using 4 cores is not the same thing as using 44 cores; modern hardware is about using tools to determine where bottlenecks lie and not about guessing. This talk dives deep under the covers of your processor and memory system to deliver top-notch performance, all while using C# and .NET. We will see multiple examples of optimizing CPU- and memory-bound algorithms that are used in real-world applications, such as finance, image and signal processing, and many others.