Speed of a Bible search

I was trying to check a recent Rebuild Bible Index command that appeared to complete successfully by running the following search. I am quite surprised at the immense length of time this took to complete (over 10 minutes). For comparison see the same search in every resource in my library (which includes the Bibles searched in the more limited search) (about 17 seconds). I thought the reason for a separate Bible Index was to make this sort of searching much faster.

Find more posts tagged with

Comments

NetworkGeek

I wonder if it's the Grid format? I did the same search as you, against Top Bibles, and it was 0.53 seconds if Verses, 27.02 seconds next if Grid.

Next I used a Collection of Bibles I have, that is 9 Bibles, and grid took 42 seconds.

Not sure how many Bibles in your Collection, but I bet it's more than 9. Try searching in Verses I bet it is much faster. Or use a smaller Collection with Grid and it will be much faster.

NetworkGeek

https://community.logos.com/discussion/comment/189784#Comment_189784

Ah I see in your graphic, your Collection has 34 Bibles, so I bet it's that + Grid = slow...

Melissa Snyder

We do plan to optimize the Grid view. I will add a link to this thread to the current case.

Mark Smith

https://community.logos.com/discussion/comment/189787#Comment_189787

Ah I see in your graphic, your Collection has 34 Bibles, so I bet it's that + Grid = slow...

I would agree that grid view must add to the rendering time, but I can't understand it contributing to the search time. It clearly must be doing so since that same set of resources searched using Verses view takes only 2.2 seconds.

That's the same time it takes to do this in Basic Search:

I am surprised that there is a ten minute penalty for running this search in Grid view. (Probably even longer in Analysis view). Other than produce the pretty colored boxes in Grid View what does Logos have to do that takes so much time? Inquiring minds want to know.

Mark Smith

https://community.logos.com/discussion/comment/189804#Comment_189804

We do plan to optimize the Grid view. I will add a link to this thread to the current case.

You answered while I was posting my reply to Dominick.

Can you or anyone explain why the actual search takes so much longer? Does the 620 second number include the time for rendering? (It doesn't seem to have anything to do with rendering in other searches and certainly isn't all the rendering that will need to be done, just one 'page' of it.) Why is grid search so slow is what I'm asking.

Melissa Snyder

https://community.logos.com/discussion/comment/189809#Comment_189809

I can't answer that question; perhaps a developer will jump in. I do know, though, that the search term can affect the length of time, as a search for 'god' on my system, in a collection of 65 Bibles, took over 14 min. However, a search for 'testimony' took only 10.19 secs -- Grid view in both instances.

Mark Smith

https://community.logos.com/discussion/comment/189811#Comment_189811

Thanks, Melissa.

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/189809#Comment_189809

Does the 620 second number include the time for rendering? (It doesn't seem to have anything to do with rendering in other searches and certainly isn't all the rendering that will need to be done, just one 'page' of it.) Why is grid search so slow is what I'm asking.

Yes and no. Drawing the "pretty colored boxes" does take a bit longer than just displaying the results of a "Verses" search, but the bulk of the search time is used to create the grid. As you search across Bibles with different versifications (see also http://blog.logos.com/archives/2007/07/books_chapters.html), a lot of time is spent in putting the different versions in the right cell on the grid. For example, if you run a Grid search in Revelation for "dragon OR sea", the grid graphically shows you which versions have verses 12:17, 12:18, and 13:1, which combine 12:17 and 12:18, and which combine 12:18 and 13:1. (Note also that the verse labels on the left change if you change the selected version. Furthermore, some data is incomplete: for example, the NIV shows a "missing" circle for Rev 12:18; this is because we haven't yet tagged it with an NIV-specific versemap marking 12:18 as being included in 13:1.)

We've added a lot of versification schemes since the grid was created, and it's quite possible that that's now exposing some inefficiencies we weren't aware of before. For now, you can speed up the grid by searching a smaller range of books (e.g., search just the Pentateuch, not the whole Bible), or by searching fewer Bibles.

Mark Smith

https://community.logos.com/discussion/comment/190432#Comment_190432

Thanks for the explanation, Bradley. I knew that those pretty colored boxes couldn't be at fault. (We certainly can't be without them.)

While I would never repeat the search I performed for this test, it does seem to take an extreme amount of time compared to doing the same search in other ways. As I pointed out a basic search finds all the same hits in a fraction of the time, so about 10 minutes is spend figuring out which part of the grid to display the hits in. That does seem like a lot and must have to do with how it is currently being done.

This is the first time I've run into this. Most searches I run are on one version so none of the 'grid' overhead comes into play. I imagine I won't do a search like this ever again except for an index test as I did this time.