Possibilities to improve iMatch performance

[suttonbg]

Dear Mario,

This request for advice somewhat follows a thread started by Carlo Didier on 7 May 2015.

My standard setup has most of iMatch (executables, cache) on an older SSD which operates as C: drive. As an experiment, I off-loaded a bunch of programs to make sufficient room to move the database folder to C: drive and noticed, not surprisingly, a hefty improvement in performance. Needless to say, a current model 1TB SSD has been ordered.

I'm doing a lot of cleaning up as my wife and I have merged our image collections, with inevitable duplication etc. I'm making extensive use of the Timeline view, filters and the Renamer, so a lot of files are being moved, renamed, re-categorised, etc. I notice that this, again not surprisingly, means a lot of read/write operations to the image files, as shown by the resource monitor. Operation of iMatch can become quite slow or stop-and-start at these times.

In rough handfuls, I have about 6TB of images, so holding those on SSDs is too expensive to contemplate. The images are held on NAS-grade HDDs in Raid 1 configuration in Windows Storage Spaces. This raises the question of whether there is any benefit from invoking read/write caches for the HDDs, perhaps either using Storage Spaces itself (eg https://www.reddit.com/r/DataHoarder/comments/lsjs5j/adding_an_ssd_to_microsoft_storage_spaces_to_act/) or using a third-party product such as Primocache?

[Mario]

Keeping the IMatch database on the fastest disk (SSD) brings the most performance.

When you mean IMatch is 'slow' while moving or copying files, keep in mind that typically 99% of that time is spent "inside" the Windows routines that perform the move and copy operations. The rest is a bit of house keeping in the database.

As usual, it would help if we can see an IMatch log file (see log file) from a typical IMatch session. Switch to debug logging via Help menu > Support for a more detailed log file. Then perform some of the tasks you consider "slow". Use Help menu > Support > Copy Logfile to save the log on disk, ZIP and attach to your reply.

This shows us what you do, how long each step takes, why IMatch is very busy etc. This will allow us to provide tips regarding performance. Generally, IMatch is very fast.

But users can shot themselves into their own foot by overly complex or broad file relation settings. By creating 'expensive' data-driven or formula-based categories etc. The log file will tell us more.