Clicking on C: will let us modify settings, so let’s select 4K blocks, set the cache size to 4096MB, enable deferred writes, then select Start Caching. That’s it. You’re now caching.
Now, we can look at the performance monitor by clicking the button in the status window:
Right after starting the cache, the hit rate is practically 0%. As the time and workload increases, the right data starts to get cached. Here is the performance monitor during an intense, protracted workload:
The particular SSD in this test, a 120GB Mushkin Chronos Deluxe, is being used as the system drive. It’s had weeks to settle in, and performance has become quite average. So before looking at benches with caching enabled, let’s get a baseline.Although this run with Anvil’s Storage Utilities is with almost incompressible data, performance is still well below where it is fresh out of box. It’s clearly in a used, steady state. So can a little caching help improve the performance?
That would be a most definite yes! The above numbers are just plain silly, showing off what 1333MHz DD3 CAS 9 can do when accelerating a modern SSD, even one suffering the effects of performance degradation. But something else is going on here — those face-melting 511,167 4K random read IOPS? They’re being CPU limited. That is, the number one performance limitation in this particular circumstance is the CPU. Even a small increase in CPU performance could greatly increase the result here.