On 2017-06-11 16:30 +0200, Toby Douglass wrote: > I wrote: > > Using the OS GCC (4.9.2, native build for i586-linux-gnu); > > WIthout, I get arch i586, tune generic. > > With, I get - ha! - "-march=native", and no tune. > > This seems odd. > > A bit more information now, pairs of command/output. > > First on Debian 8, the OS provided 4.9.2 on my normal x86_64 machine. > > gcc -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-E' '-v' '-mtune=generic' '-march=x86-64' > > gcc -march=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-march=native' '-E' '-v' > > gcc -mtune=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-mtune=native' '-E' '-v' '-march=x86-64' > > gcc -march=native -mtune=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-march=native' '-mtune=native' '-E' '-v' Oh no Toby... I didn't mean COLLECT_GCC_OPTIONS. Its -march is same as your input (or defaulted). I mean: gcc -march=native -E -x c /dev/null -v 2>&1 | grep cc1 | fmt /usr/lib64/gcc/x86_64-pc-linux-gnu/7.1.0/cc1 -E -quiet -v /dev/null *-march=ivybridge* -mmmx -mno-3dnow -msse -msse2 -msse3 -mssse3 -mno-sse4a -mcx16 -msahf -mno-movbe -mno-aes -mno-sha -mpclmul -mpopcnt -mno-abm -mno-lwp -mno-fma -mno-fma4 -mno-xop -mno-bmi -mno-sgx -mno-bmi2 -mno-tbm -mavx -mno-avx2 -msse4.2 -msse4.1 -mno-lzcnt -mno-rtm -mno-hle -mno-rdrnd -mf16c -mfsgsbase -mno-rdseed -mno-prfchw -mno-adx -mfxsr -mxsave -mxsaveopt -mno-avx512f -mno-avx512er -mno-avx512cd -mno-avx512pf -mno-prefetchwt1 -mno-clflushopt -mno-xsavec -mno-xsaves -mno-avx512dq -mno-avx512bw -mno-avx512vl -mno-avx512ifma -mno-avx512vbmi -mno-avx5124fmaps -mno-avx5124vnniw -mno-clwb -mno-mwaitx -mno-clzero -mno-pku -mno-rdpid --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=3072 -mtune=ivybridge So we can see GCC guessed "ivybridge" on my laptop. In fact it's correct. W/o "-march=native": gcc -E -x c /dev/null -v 2>&1 | grep cc1 | fmt /usr/lib64/gcc/x86_64-pc-linux-gnu/7.1.0/cc1 -E -quiet -v /dev/null -mtune=generic -march=x86-64 So the default is "x86-64". > Interetin', eh? tune is doing something, even though arch isn't. -mtune is not doing anything. With only -mtune (no -march) GCC just use the default -march (configure option --with-arch=). For example, we have a Ubuntu server with GCC 4.9 from APT. cc -v 2>&1 | fmt | grep with-arch-32 --enable-objc-gc --enable-multiarch --disable-werror *--with-arch-32=i686* So... cc -m32 -E -x c /dev/null -v 2>&1 | grep cc1 | fmt /usr/lib/gcc/x86_64-linux-gnu/4.8/cc1 -E -quiet -v -imultilib 32 -imultiarch i386-linux-gnu /dev/null -m32 -mtune=generic *-march=i686* -fstack-protector -Wformat -Wformat-security > Going back to the version 1 Minnowboard (Atom E640), Debian 8, the OS > provided GCC 4.9.2; > > gcc -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-E' '-v' '-mtune=generic' '-march=i586' > > gcc -march=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-march=native' '-E' '-v' > > gcc -mtune=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-mtune=native' '-E' '-v' '-march=i586' > > gcc -march=native -mtune=native -E -x c /dev/null -v > COLLECT_GCC_OPTIONS='-march=native' '-mtune=native' '-E' '-v' > > Superficially at least, -march is looking a bit messed up. However, > it's more likely my expectation of what it is does is amiss =-) -- Xi Ruoyao <ryxi@xxxxxxxxxxxxxxxxx> School of Aerospace Science and Technology, Xidian University
