aaupov wrote: > prefer to use the perf script format directly, as it offers advantages when > running perf2bolt multiple times for dependent DSOs
I'm adding support for emitting pre-aggregated profile (#199465) which can be extended to cover DSOs. This way, you could use one input profile to optimize main binary and DSOs (with or without converting to fdata/yaml – #199322 drop the warning that you shouldn't use perf/pre-aggregated for optimization directly). This is what we use internally. https://github.com/llvm/llvm-project/pull/199324 _______________________________________________ llvm-branch-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
