Data parallel analsysis
konichuvak last edited by
Anyone figured a clean way to attach individual analyzer per datafeed? The purpose is to compare performance of the strategy using build-in analyzers (return, sharpe, drawdown, etc) across multiple venues. I am aware that analyzers are meant to operate on strategies, however I was wondering whether anyone came up with a reusable pattern to write analyzers that operate on datafeeds.
For now, I am simply running an individual backtest per datafeed in a separate process, which allows for such parallelism but introduces some overhead.