LLVM Lands Support For The Marvell ThunderX3
In April I wrote about the ThunderX3 support for the GCC compiler getting squared away while this month the LLVM compiler support is ready.
Being merged on Wednesday was the initial ThunderX3 (ThunderX3T110) support.
With LLVM Git or what will become LLVM 11.0, the -mcpu=thunderx3t110 option is supported for those wanting to target this new Marvell/Cavium server CPU. This flag enables ARMv8.3-A plus enabling all of the relevant extra instructions supported by this processor as well as being aware of the ThunderX3 attributes.