Hi, I'm looking for the definition of 3 columns in the ExecutionLog table. 1. TimeDataRetrieval 2. TimeProcessing 3. TimeRendering What each one of the times include? and in what is the unit of this numbers (miliseconds)? Thanks, Assaf

They are in milliseconds. TimeDataRetrieval is the amount of time spent running the datasources for your report. TimeProcessing is the time spent processing the report. TimeRendering is the time spent generating the output format that you requested (for example, HTML). Note that these are not necessarily additive to the total time required to run the whole report because some activities -- such as timeprocessing and timedataretrieval can actually be interleaved.

Hi, i want ask question about the columns which are the TimeDataRetrieval and TimeDataProcessing.If i am a reporting service manager, which onewas the most important critier for me , retrieval or processing .Could you help about this subject.thanks...

If you are concerned about optimizing performance, then take the largest one and try to decrease it TimeDataRetrieval is the time that RS is spending retrieving results from the data source. If this number is really large that usually indicates a poorly performing data source or your are trying to pull too much data into the report server. You should consider optimizing your queries (rewrite, add indexes, etc...) or retrieving less data potentially by parameterizing them. Large TimeDataProcessing could again indicate that you are trying to pull too much data into the report server, or that your grouping/sorting/aggregation expressions in the RDL are suboptimal or too complex. You should try reducing the complexity of the report. Hope this helps.

If I run web performance tests on any of my reports, the time reported is always larger than the sum of these times. Shouldn't interleaving improve performance (in most cases) ? I'm running R2 and the server is on localhost. Any ideas on why I observe this behavior ?