Updated Compiling from source (markdown)

JayDDee
2025-11-03 11:27:20 -05:00
parent f32bd2022a
commit 180ea06bd6

@@ -158,6 +158,11 @@ N is the number of threads for the compiler to use. Use `-j $(nproc)` to use all
Some newer CPU architectures may not be defined for `-march=native` in older versions of the compiler and may not recognize new features using the default build script. It is often possible to add the missing features to create an optimum build by manually specifying options. Some newer CPU architectures may not be defined for `-march=native` in older versions of the compiler and may not recognize new features using the default build script. It is often possible to add the missing features to create an optimum build by manually specifying options.
MacOS & CLang have difficulties with native compile on Apple M series CPUs, crypto may be missing. All M CPUs have crypto (AES & SHA256) therefore the default build may not be optimum. To ensure optimum build use the following instead of `-march=native`.
M1: `-march=armv8.4-a+crypto+sha3`
M2,M3: `-march=armv8.6-a+crypto+sha3`
M4,M5: `-march=armv9.2-a+crypto+sha3`
Many examples can be found in file build-allarch.sh for x86_64 CPUs, and in armbuild-all.sh for ARM64 CPUs. Many examples can be found in file build-allarch.sh for x86_64 CPUs, and in armbuild-all.sh for ARM64 CPUs.
Windows on ARM64 may crash with a segfault. See Windows section above. YMMV Windows on ARM64 may crash with a segfault. See Windows section above. YMMV