2 Replies Latest reply: Apr 12, 2011 2:11 PM by cliffwong RSS

    High CPU Utilisation

      Server: 8 core Blade with 64Gb RAM

      2 transactional tables (1 with 8m and the other with 600k records) and 1 calendar island, with only 1 graph displayed.

      Calculating a ratio between two fields from the tables and displaying in a graph with time from the calendar at the x-axis, QV manages to utilise all cores (>90%) and about 3Gb RAM for about 10 seconds, when refreshing or filtering on the graph.

      Any reason why CPU was fully utilised compared to RAM?

        • High CPU Utilisation
          Daniel Rozental

          You should probably try not using an island calendar.

          You could also try posting a sample app or attaching the model & expression so it will be easier to help you.

            • High CPU Utilisation

              Attached is a file utilised to test file structure.

              I am trying to calculate 2 loss ratio graphs (as highlighted in red) by joining the fact tables with the ProductNumber. Have not added the date field in the link as there are multiple date fields in the fact tables.

              Essentially, what I am trying to replicate from a traditional BI environment is a nested query, where the Claim value (grouped by Date, Product Number and restricted by Claim Attributes) is compared to the Premium Value from the same set of Product Numbers and Date, to calculate ratio. This is normally a 2-3 step process, but not sure how to replicate these processes within QV.

              As correctly noted, volume of entries prohibits use of Island Calendar.