When I search for files directly in Everything, the speed is very fast, and I basically don't feel the delay. When I use dopus to search for Everything, I have to wait for a long time to get results. For example, input folder: filter file folder, and wait about 10s in dopus. dopus13 is not directly calling Everything. How is the speed still inferior to Everything?
Why is the speed difference between direct Everything search and calling Everything to display resul
Opus will always be slower than Everything for having to transfer and translate the information. However, 10 seconds sounds like a little too much. For me it's pretty much instant upon pressing Enter, otherwise taking no more than 1 second (there is a little delay in Opus by default, which will be configurable in future version 13.3.1).
Same here.. if I use the dedicated search field in DO in Everything mode, it is a night and day difference compared to the experience when using the Everything GUI directly. It is a bummer! o)
I have 7 millions indexed objects and the Everything GUI searches through them instantly while typing, no delay. When in DO on the other hand, "everything" locks up (literally! o).
I wonder if it's the API of Everything v1.5a, it is still in alpha stage, but given the results come up that fast in Everything, I do expect DO to be slower, but not that slow?! In DO I find the E-search is actually not usable (on my system at least).
The first search from DO is kinda "ok" regarding response time, but editing the search term basically locks up the search completely, the Everything GUI goes into "does not respond"-mode.
Time will fix this hopefully! o)
I would like the Everything search in DO to be very similar to what the native Everything GUI offers of course, and I think the world needs to be able to use the quick filter bar in DO directly to execute the Everything query - using the dedicated search field does not help "immersion". o)
First world problems compared to file managers on Linux of course, everyone should try a Linux file manager from time to time - just to come back to DO and appreciate it even more. o)
Opus will never be as fast as using the Everything GUI directly.
So the Everything GUI is not using the same API which DO is using?
Or what other reason is behind your statement?
Technically, I would assume it should not matter where the query is entered, as long as there is no additional network distance to overcome, results could be equally fast.
Propagating the file display in DO maybe takes away some of the speed, but it's quite fast also, but surely not as fast as the more simple list in Everything.
Yes, yes.. some losses are in there, no doubt about it, but for a result list of around 100 items, which is instantly there in Everything GUI, but takes 10-20 seconds to show up in DO.. there must be something else going on than just copying less than a megabyte of data around in memory and building a list of items?!
My "daily files" folder has 3000+ files in it, a new tab with that path opens in less than a second including all the file attributes and colouring (in the visible part).
There is a performance gap there, which should not be there. It's rather looking like some bug or hick up in the API for me when looking at these numbers. Well, maybe the cause for this huge difference can be nailed down at some point, thank you! o)
Something wrong on your machine maybe, it's nowhere near that slow here. Maybe try rebuilding the index.
When searching for a file, if the search results are few, the two search methods basically do not feel the difference. If there are many search results, such as more than 100,000, the difference is particularly obvious.
I suspect that the two do not use the same api, or everything has its own direct method, which is not open.
Thanks for all your hard work.
Just for comparison sake, if you do a global search on your machine for "*.exe" how long does it take for the results to be displayed and how many results are there?
It tops out at 10,000 due to the default limit, but that takes around a second.
dopus does limit the number of displays, and there is not much difference between the two in this order of magnitude. When I set the display quantity to one million, the difference is very obvious. Guess if there is any mechanism within everything that can be optimized
My Everything experience in DO:
(please note weird blue circle behaviour at 0:50, first query returns at 1:55, nearly 2 minutes):
My Everything experience using the Everything client:
Night and day, or it's more like there are universes in between! Updated the Everything.exe to latest alpha before. Not sure what's going on, but as you can see, it's not really usable in DO for me.
I made a quick test here (~1,200,000 entries) using the + search (global): the first time it took about three or four seconds, but after that it was almost instantly (~1 second).
Have you tried a non-alpha version (i.e. 1.4)?
It could also be tied to what Everything is configured to index.
No, after looking at the improvements in v1.5, my impression is that maintaining or setting up v1.4 is not worth it, even if it means not having a working DO integration. At some point, v1.5 will be official and DO will probably work fine with it (for me too), I will wait. I'm not a heavy Everything user yet.
I also noticed that getting size and item count information from E is nice at first, but many times these values do not update fast enough for me and I cannot force F5 things anymore after deleting/adding to a folder, which makes this feature nice in theory but it does not seems to work for me in day to day use.
So, whatever the problem with the query is, the DO integration does not work for me anyway it seems (no offense). I might look into it again at some point in the future, until then, the native Everything client does the job. I just wanted to agree to the thread opener and share my findings.
Do you think? But how, if any response in the Everything client is basically instant?
We'll never know if the problem is related to that unless you check, since we aren't seeing any issues on our machines. AFAIK 1.5 has been in alpha for a long time and may stay in alpha for a long time too (although I may be wrong!).
Opus uses Everything's index differently to how Everything itself uses its own index.
For example, if Everything isn't providing certain sizes/attributes then Opus may be forced to fall back on the old non-indexed way of calculating things.
It may turn out to be unrelated to the index options or Everything version, but the only way to find out is to try them.
Do you notice that the Everything app goes into a "not responding" state when Dopus is trying to display the search results?
It's always been like this
You're all insisting on using the 1.5a version even when we tell you it's an alpha and we haven't tested with it. These are the sort of weird effects that I would not be surprised to see in an alpha version.
1 Like