[DO11 bug] slow settings search

I have quite fast PC, i7 CPU clocked on 4.2GHz; after entering preferences filter string for the first time, it takes about 10 seconds for search to be finished. Even though it's quite low priority problem, I'd say something is not right here.

P.S. I understand that you probably instantiate all pages on the first search, but 10 seconds is still quite a lot of time.

It doesn't take that long here, but I guess might if the machine is running/paging slowly. Or maybe if one of the configuration pages takes an extraordinarily long time to initialise for some reason (like if you have a million Favorites or something).

It only happens the first time you search, should only take a few seconds on most machines, and we don't consider it a bug.

Hm. CPU usage is <5% before doing search, I have 32Gb of RAM, so there is no paging.

Also, I'd say it's not a problem in one page because if, for example, string appears on ten pages, similar amount of time is needed for each one of them to appear.

But I agree - if this is not a commonly reported problem, skip it for now.

How do you know?

Each page requires different work to initialise. The actual string searching is the trivial part.

Try closing the Prefs dialog, then open it up and go through each page one by one. Maybe one of them will take a long time to appear the first time.

Since 16Gb is mostly free. :smiley: Windows had some strange swapping behaviour in the past, but, the main trigger for that was actually the lack of memory and misinterpretation of counters.

This can easily be verified - process working sets are not trimmed (they are at the level of private memory usage) and RAMMap utility shows that cache holds files that were last accessed days ago.

Yup, I understand that.

When you divide, you get about 100ms per page which seems more or less reasonable, but 10s all together still seems excessive.

Haven't found one - all of them seem to open instantly.

Anyhow, as I said, I've reported because I've just noticed that and I was in a reporting mode; it doesn't bother me much. But if you'd like to investigate it further, I'd be happy to assist (especially if you implement some of my suggestions :grin: ).

Paging isn't just a factor of free memory. If something has never been loaded into memory until it is needed, paging it in the first time can still be slow if the disk is tied up.

(Not saying that's definitely what is happening, though.)