[flang-commits] [clang] [flang] [llvm] [flang][cmake][perf-training] Optimize flang with PGO and BOLT (PR #197947)

[Paul Osmialowski via flang-commits]

pawosm-arm wrote:

I've rebased this one too. Apparently, the BOLT buffer size needs to be extended here too in order to make it complete the full round through the perf-training cycle. I've also introduced the changes needed to keep it in sync with the competing PR #198863

https://github.com/llvm/llvm-project/pull/197947