I know for most people this kind of stuff probably doesn't matter but I try to squeeze every bit out of A-parser and want to make sure I understand about using multiple cores. If running single task there is no benefit in speed from multiple cores but if running multiple tasks at once, more cores will help speed up job. What I am not clear on: 1) If I have 4 tasks at once is just a single core used per task? Is there any benefit to set number of cores higher than the number of running tasks? 2) For processing cores: "Multiple cores are used only for filtering, Results Constructor, Parse custom result". Does this mean if I am just downloading the full html page with no custom parsing or filtering there is no benefit to allocating multiple cores to processing? If I have 4 tasks running that just downloads raw data would 4 cores for task and 4 cores for processing be optimal OR 4 cores (or more) for task and just the minimum 1 core for processing be optimal? 3) When using Redis api which by default parses all possible data available in a parser (Google parser all elements are parsed) Does it matter how many cores are allocated to processing the results? If I had 3 redis tasks running Google Parser, what would be the optimal core allocation settings? For proxy strategy settings the documentation says: "Allows you to manage the proxy selection strategy when using sessions:" Does this mean sessions like Google Parser uses sessions and if for some reason you don't use sessions there is no benefit to "retain the proxy from a successful request for the next request" and random will be fine/possibly better? Thanks.
Yes - one task uses up to 1 core for data processing, but due to the nature of the JS garbage collector, a single task may still use 2-3 cores. There’s no benefit in allocating more cores than the number of running tasks. Yes, there are no benefits in this case. Moreover, we’ll remove this setting soon. 3 tasks = 3 cores for task processing. There’s no definitive answer about the proxy strategy unless you test both approaches.
