perf-staging compiled with -O0

[mmeeks]

It seem we're generating bogus profiles on staging-perf because the code is being compiled with -O0 - and this generates -amazingly- rubbish code; eg. the SIMD Delta behavior is slower than the CPU version, despite being 2x as fast when compiled properly.

This distorts everything we do when it comes to optimization; it would be -really- good to fix that ASAP :-)

proof:

strings /usr/bin/coolforkit | grep arch=x86
GNU C11 7.5.0 -mtune=generic -march=x86-64 -g -g -O2 -std=gnu11 -fgnu89-inline -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fmerge-all-constants -frounding-math -fstack-protector-strong -fpatchable-function-entry=16,14 -fmath-errno -fno-stack-protector -ftls-model=initial-exec -fPIE
GNU C++14 7.5.0 -mtune=generic -march=x86-64 -g -O0 -std=c++1z
GNU C11 7.5.0 -mavx2 -mtune=generic -march=x86-64 -g -O0
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Can we check the product builds too to ensure they are optimized as expected ? =)

@caolanm might be interested too ...

[caolanm]

strings now has:

strings /usr/bin/coolforkit | grep arch=x86
GNU C11 7.5.0 -mtune=generic -march=x86-64 -g -g -O2 -std=gnu11 -fgnu89-inline -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fmerge-all-constants -frounding-math -fstack-protector-strong -fpatchable-function-entry=16,14 -fmath-errno -fno-stack-protector -ftls-model=initial-exec -fPIE
GNU C++14 7.5.0 -mtune=generic -march=x86-64 -g -O2 -std=c++1z
GNU C11 7.5.0 -mavx2 -mtune=generic -march=x86-64 -g -O2
GNU C++14 7.5.0 -mtune=generic -march=x86-64 -g -O2 -std=gnu++14 -ffunction-sections -fdata-sections -frandom-seed=ops.lo -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fPIC
GNU C++14 7.5.0 -mtune=generic -march=x86-64 -g -O2 -std=gnu++14 -ffunction-sections -fdata-sections -frandom-seed=path.lo -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fPIC
GNU C++14 7.5.0 -mtune=generic -march=x86-64 -g -O2 -std=gnu++14 -ffunction-sections -fdata-sections -frandom-seed=dir.lo -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fPIC
GNU C11 7.5.0 -mlong-double-80 -mtune=generic -march=x86-64 -g -g -g -O2 -O2 -O2 -fmessage-length=0 -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fbuilding-libgcc -fno-stack-protector -fpic -fvisibility=hidden
GNU C11 7.5.0 -mtune=generic -march=x86-64 -g -g -O2 -std=gnu11 -fgnu89-inline -fmessage-length=0 -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -fmerge-all-constants -frounding-math -fstack-protector-strong -fpatchable-function-entry=16,14 -fmath-errno -fno-stack-protector -fPIC -fstack-protector-strong -ftls-model=initial-exec

so I imagine we can now close this