improved semantic checking performance for project hierarchies
Submitted by Kristof Szabados
Link to original bug (#568445)
Description
At the last colorful day, we have seen that running the semantic checking in parallel mode can be ~3,4* faster on a 4 core and ~10* faster on a 28 core machine compared to the single mode. But this does not apply to complex systems, that use project hierarchies. Where the issue is, that each project in such a dependency hierarchy has to be analyzed independently today to make sure.
Which was not a problem when we did not have parallel semantic checking, but now could be further improved. If the checking would collect all modules from all reachable projects and analyze the min parallel + as a separate step do the dependency check, we could get the functionally same result, but much faster coming from the opportunity created by several modules being present for parallel processing at the same time.
Version: 7.1.0