Profiling samples on the eventlog are all given very similar timestamp
On a simple test program is seems that events are not properly added to the eventlog buffer. I would at least like to understand why this is.
Starting from !1340 (closed)
module Main where
main = print $ foldl (+) 0 [1..10000000]
ghc Run.hs -prof
./Run +RTS -hb -l-au
The samples in the resulting eventlog are all too close together. It seems the time comes from when the buffer is flushed rather than when the events are added to the buffer.
https://gist.github.com/mpickering/6d4a9e47047aa41f452930121688ab27
The .hp
file looks fine.