For anyone else looking for a solution here, the GDAL command worked pretty quickly.

Looks like this issue is documented in a few places: issues.qgis.org/issues/15292 issues.qgis.org/issues/18406 issues.qgis.org/issues/20008 But this one seems the closest to what I describe: issues.qgis.org/issues/19771 Doesn't look like there is a real fix at the moment.

Thanks for the reply. I think that something different is going on in my case since I used the main window to do the edits instead of the attribute table. When I did try via the attribute table, I experienced the same general issues you describe here (high cpu load), but mine did not finish in a reasonable time period (~30 minutes) so I went with the main window route, leading me to my description of the problem above. I will see what I can do about making a bug report, but will probably just try to do it in GDAL for the time being.

I don't think that is the case. I just got to my computer after letting the save operation run all night, and it was STILL using up my hard drive read/write. What I don't get is that the file size was only about .3 gb (for all of the files), so obviously it wasn't actually writing the whole time.

Yeah, I meant to be filtering TOA reflectance, so the values I was using didn't make sense. The solution ended up being to add a property of the mean band value in my area of interest for each image in the collection, then filter on that. Thanks!