How big should cache size be?

Cache size is a very important cache parameter. It influences the cache performance more directly than perhaps any other cache parameter. The larger the cache, the more data can be cached, and the better the cache performance. When the cache is full of data, some old cached data is discarded according to the cache replacement algorithm to make room for the newly requested data. If Windows later requests to read the discarded data, the cache needs to read the data from the hard disk again. Obviously, a small cache increases the frequency of data discards, thereby reducing cache performance.

Of course, from an economic point of view, a larger cache means more capital investment, so you need to choose an appropriate cache size according to your budget and the amount of data on the cached disk. When using a cache to speed up disk reads, it is recommended that the cache size be as large or larger as the amount of frequently accessed data (aka hot data) on the cached disk. If you are not sure about the amount of hot data, it is recommended that the cache size be at least 10% of the total data size, preferably 20% or more. You can also adjust the cache size later based on the value of "Unused Cache" in the statistics panel. If there is a lot of unused cache, you can reduce the cache size appropriately.

Also, when setting the L1 cache size, not all memory can be used as L1 cache, but enough memory should be reserved for Windows and other applications. You can check how much free memory is left after all frequently used programs are running, and then reserve at least 1GB (preferably 2GB or more) of memory from the free memory. The remaining free memory can be considered for use as L1 cache.