Closed jrudolph closed 8 years ago
@brendangregg thanks for testing the new functionality. The size increase of the map size is in the range I also observed. It would now probably make sense to start building something as outlined in #27 where the load for the target process is spread over its run time.
tested (with @tjake's flame graph changes), and seems to work!
This was a pretty intense and complex production workload; the perf.map file went from 8 Mbytes to 250 Mbytes with unfoldall. :)