There's more to experience when you log in!
Does the version 17 of MapInfo Pro utilizes all the processors? Any control on the number of processors that can be used?
Similar til v16.0, MapInfo Pro v17.0 can take advantage of multiple threads.
Some of the Object Processing methods and many of the Raster Processing tools take advantage of multiple cores.
On the PRO tab, Options, Performance, you can manage how many of your processors you want MapInfo Pro to use:
Notice how the levels match with percentage of your processors.
Is there much difference between the settings for "Aggressive" and "Full"?
25% difference or 4 instead of only 3 cores.
I would say if you are doing some heavy object processing, it would speed it up slightly. I really haven't testing results.
But what's more important, if you want to use your computer for other work while Pro is doing the processing, you might want to limit the Concurrency Level to Intermediate or Aggressive.
I just gave this a try using the Buffer operation that also supports multiple threads. The table I used had 560 000 records.
I used a command like this:
Create Object As Buffer From VEJMIDTE
Width 250 Units "m" Type Cartesian Resolution 100
Into Table BufferTest Group by Rowid
The last line controls the concurrency or number of threads where None means "use just one thread".
Here are my not very precise findings:
None, one thread:
I stopped this after three 2:40 minutes. Only 2% of the objects had been buffered. If the speed would be steady it would take 2 hours to buffer the entire table.
Moderate: 2:24 for all 100%
Intermediate: 2:08 for all 100%
Aggressive: 2:03 for all 100%
All/Full: 2:02 for all 100%
So for buffers, you might just settle with Intermediate as the performance doesn't seem to improve much after that. That's on my computer with 8 logical processors. I'm writing to a standard HDD so writing to a SSD might improve the performance more.
Is there a list of which operations in Mapinfo take advantage of multi-core operation?
To answer your question on which operations can utilize multiple cores, here is what I found in the Help documentation under the topic Setting Your Concurrency Preference.
"This improves the processing time when buffering an object in a table or selection, and with overlay operations (such as Split Target, Erase Target, Erase Outside Target, Split Target using Polyline, and Overlay Nodes onto Target)."