"In memory" is a hot topic in the database world at present, but how do modern servers CPUs utilise memory, does the story end with main memory ?, what about NUMA and the memory hierarchy on the actual CPU. Do memory access patterns matter ? does the CPU socket certain workloads are executed on matter and how can all of this be leveraged in the database engine to our benefit. All these answers and more will be covered at level 400 including, large memory pages, spinlocks, optimising hash joins for leveraging the CPU cache, the OLTP database engine and the LMAX queuing pattern. During this journey everything a SQL Server professional needs to know and should know about memory will be covered, along with deep insights into the database engine, CPU architectures and the use of windows performance toolkit to quantify the performance related behaviour of the database engine.
Presented by Chris Adkin at SQLBits XV