This issue tracks the work associated with identifying and solving the discrepancy between Windows Intel vs. Ubuntu Intel and Windows Intel vs. Ubuntu ARM based on the GCPerfSim Scenarios.
The general themes in terms of the results are:
Windows Intel based traces exhibited a higher execution time than their Ubuntu based counterparts.
% Time Spent in GC for Linux based runs is, in general, much higher than the Windows based counterparts.
Results on the ASPNet Machines are far less volatile (minus a few outliers sprinkled here and there) than those on our GC machines implying a need to establish some protocols such as logging other users off before running any of the tests.
Tagging subscribers to this area: @dotnet/gc
See info in area-owners.md if you want to be subscribed.
Issue Details
## Summary
This issue tracks the work associated with identifying and solving the discrepancy between Windows Intel vs. Ubuntu Intel and Windows Intel vs. Ubuntu ARM based on the GCPerfSim Scenarios.
The general themes in terms of the results are:
1. Windows Intel based traces exhibited a higher execution time than their Ubuntu based counterparts.
2. % Time Spent in GC for Linux based runs is, in general, much higher than the Windows based counterparts.
3. Results on the ASPNet Machines are far less volatile (minus a few outliers sprinkled here and there) than those on GCPerfCl implying a need to establish some protocols such as logging other users off before running any of the tests.
Summary
This issue tracks the work associated with identifying and solving the discrepancy between Windows Intel vs. Ubuntu Intel and Windows Intel vs. Ubuntu ARM based on the GCPerfSim Scenarios.
The general themes in terms of the results are: