This problem with xactlog is now fixed. Some of the arithmetic calculations with the event times was overflowing 64 bit ints, which was a problem for runs with long execution time.
By the way, as a general rule to get the most accurate profiles, you should avoid generating activity logs from a REPL interaction. You can do this with the -e command line option.
Marc
On Oct 31, 2016, at 4:25 PM, Marc Feeley feeley@iro.umontreal.ca wrote:
By the way, your screendump is strange… the middle part, the “percent profile”, should be a summation (put on a scale of 100%) of the activity bars above (one per processor). But the red areas (for the GC) don’t line up… Could you try to run xactlog with the -us or -ms option (to scale to microseconds or milliseconds)?