You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Looking at a few other packages I noticed a 10-15% slowdown between 0.4 and master. However, DistributedArrays is showing a slowdown for the entire test suite run of around 20-30%. The underlying causes for this may very well be in Base. This issue has been created to help track down specific instances of slowdown so that they may be addressed better in Base.
The text was updated successfully, but these errors were encountered:
Are these actual performance regression or is this the penalty coming in from longer compile times + closures? We should probably come up with a set of benchmarks that we can test with BenchmarkTools
Anecdotal at this point. Considering longer compile times are the same across packages and that closures have been optimized in 0.5, comparing, for example, PR histories of DataFrames and ProtoBuf, they have seen a slowdown of approx. 10% while it is considerably more here.
amitmurthy
changed the title
Performance regression on master (test suite run timings)
Track performance regressions.Implement set of benchmarks.
Aug 3, 2016
Looking at a few other packages I noticed a 10-15% slowdown between 0.4 and master. However, DistributedArrays is showing a slowdown for the entire test suite run of around 20-30%. The underlying causes for this may very well be in Base. This issue has been created to help track down specific instances of slowdown so that they may be addressed better in Base.
The text was updated successfully, but these errors were encountered: