SD812F 3BDH000014R1 Learn about the cache Compensate Processor

Cache performance measurements become important in an era when the speed gap between memory performance and processor performance is growing exponentially. Caching was introduced to reduce this speed gap. Therefore, understanding the cache’s ability to bridge the processor and memory speed gap becomes important, especially in high-performance systems. The cache hit ratio and cache miss hit ratio play an important role in determining this performance. In order to improve the cache performance, it is necessary to reduce the missed hit ratio. Reducing the access time to the cache also helps improve its performance.

CPU stop
The time it takes to get a cache line out of memory (read latent due to cache misses) is important because the CPU will have nothing to do while waiting for a cache line. When a CPU reaches this state, it is said to be stalled. As the CPU becomes faster than main memory, pauses due to missed cache hits replace more potential computation; Modern cpus can execute hundreds of instructions in the time it takes to extract a cache line from main memory.

During this time, various techniques were employed to keep the CPU busy, including out-of-order execution where the CPU tried to execute separate instructions after waiting for instructions to cache missed data. Another technique used by many processors is simultaneous multithreading (SMT), which allows another thread to use the CPU core while the first thread waits for needed CPU resources to become available.