A Performance Tester/ Troubleshooter, needs to understand that the way an application uses memory can affect the overall application/system performance. The presentation will cover memory concepts like the cache and NUMA along with coding examples showing good and bad cache usage. Furthermore, the presentation will explain how threaded code can be further impacted by false sharing, memory allocation contention and NUMA memory placement. Two tools (Intel Amplifier and ThreadSpotter) that I’ve used to help uncover these types of performance issues will also be discussed.
What is the memory cache across multiple architectures?
What is NUMA?
A Tester will understand the knowledge to work with developers on memory related performance issues.
Exposure to 2 tools that can help uncover performance issues.