 |
 |
Someone requested a response to this question I had earlier (below). I'm guessing a bunch of people are looking to cache some of that 2016 imagery Emily announced last week, if they haven't done so yet. I was told that it N can be set to the number of logical processors +1 to get the instance setting and that licensing does not matter in this case. So I'm going to set it to 25 in my case. If that's correct, performance should be way better than it used to be when I ran it at 5. I think Esri should clarify this when they say "cores" in their documentation. Additionally, http://wiki.gis.com/wiki/index.php/Server_Software_Performance in the section "Batch process service configuration (geoprocessing services)" goes into a bit more detail on tuning both the minimum and maximum instance settings. Nick Good Morning Arc Users, I'd like to optimize the # of instances setting for the CachingTools service. ArcGIS Help says, "A good starting point is to allow a maximum of n + 1 instances, where n is the number of CPU cores on a single machine in your cluster." I have two questions. Is that the number of cores you have licensed by Server? So if I have a 4-core license, should I set that to 5 to start? I have more cores than that, so I can go higher. If I can set that higher, I have two Xeon 6-core processors in the machine. A Xeon has double the amount of logical processors, so each has 6 cores and 12 logical processors. Can I set that "n" to be the total number of logical processors or would it be limited to the # of cores? I'm not sure if the answer here is 5, 13 (6x2+1) or 25 (12x2+1). There would obviously be a big difference in caching performance and testing and comparing these could tie up my server for a couple weeks. I appreciate all responses. Thanks, Nicholas A. McNamara, GISP GIS Coordinator South Norwalk Electric and Water Phone: (203) 866-4446 ext. 2035 ------------------------------------------------------------------------- This list (CTGIS-L) is an unmoderated discussion list for all CT GIS Users. If you no longer wish to receive e-mail from this list, you can remove yourself by going to http://listserv.uconn.edu/ctgis-l.html