Open apankov opened 7 years ago
Hmmm, I struggle to think of a scenario where this code path would make any real difference to performance considering this operation will be I/O bound.
Have you measured performance here – or was this a pre-emptive change?
That's rather pre-emptive strategy. I foresee my case where I'd like to collect lots of data per shot. I had very preliminary tests on big runs and got empirical understanding of performance penalties. I understand that's not the way how to measure performance, but anyway.
Hey Michael, what do you think about micro-optimization? Thanks