Oliver Stöneberg

Given the improvement this change provides and how it affects the CI (we are actually getting even slower and not having this applied makes it harder to pinpoint the other...

Having this done in incremental commits would also make it possible to bisect differences in behavior/performance this might have based on the stage this cache is generated. Especially after I...

I dropped the changes to generate these beforehand as it was too unspecific. The performance gains are just too big to delay this any longer. I will file a ticket...

> As far as I know we have had a focus on Linux and I fear it might have hurt Windows experience. Official Windows release binaries are still built without...

Please also run Linux with `--executor=thread` (quite some data has been added to the IPC recently) and Windows with Boost.

I did some short test on my system with a smaller file and I did not see any differences between using thread or not. The process executor code also has...

> Does this indicate that if we implement a process executor in windows we might get 10x performance boost? No - it indicates there is an concurrency issue when using...

Do you actually have 16 thread available on your system? If you specify more threads than available the analysis will obviously be much slower since you are overloading the system....

> Do you actually have 16 thread available on your system? If you specify more threads than available the analysis will obviously be much slower since you are overloading the...

I am not able to reproduce this locally. I also thought it might have been caused by your CPU potentially having P- and E-cores but it not occurring with processes...