I have more questions. When I like a program so much I like to know everything I can, without of course asking you to give away trade secrets.
1. Do you know of which deferred write mode yields best overall performance? Have you done any study? It seems like native would just be best from the nature of the way it works. It is the simplest and straight to the point of writing the data.
2. Averaging. When does the writing actually commence? Is it when the interval occurs or does it begin to write immediately to try to complete the write by the time of the next interval? Or is it just a slower background priority write? Or is there a calculation based upon the amount of data to write?
2 Questions about write mode
Re: 2 Questions about write mode
Each mode has its best use scenarios. However, as you said, the native mode is the simplest and this is the default setting. The Intelligent/Idle-Flush/Buffer mode is based on the native mode and it also flushes certain amount of deferred write-data to disks when Windows is idle, thus making enough free space for future writing. You may have known that when cache space is full of deferred write-data, if at that time new writing requests come, PrimoCache will hold requests until there are enough free cache space made by "urgent writes". Obviously, during "urgent writes" the write performance becomes bad. The Intelligent/Idle-Flush/Buffer mode is designed to address this problem by making certain amount of free space in advance.
The average mode is designed for heavy writing scenarios. It will calculate the amount of data to write during certain period and decide the writing way. The purpose of this mode is to maximally use the underlying disk's write ability and handle more write IOs.RobF99 wrote: ↑Thu Oct 31, 2019 7:32 am Averaging. When does the writing actually commence? Is it when the interval occurs or does it begin to write immediately to try to complete the write by the time of the next interval? Or is it just a slower background priority write? Or is there a calculation based upon the amount of data to write?