I did a quick check on our x86_64 builders in both hardware and virtualized.

The flags which the virtualized /proc/cpuinfo say is

flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm cpuid_fault epb invpcid_single pti ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm xsaveopt cqm_llc cqm_occup_llc dtherm ida arat pln pts

And the ones in the real hardware are:

flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx rdtscp lm constant_tsc rep_good nopl xtopology cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm invpcid_single pti ibrs ibpb fsgsbase bmi1 avx2 smep bmi2 erms invpcid xsaveopt arat

Both say they have avx2 but I do not know if that is a valid test for you or not.

@smooge This is a bit of a sore issue. CPUID bits show only availability. applications also have to check using XGETBV whether the feature is not just available, but also usable (meaning that the firmware, hypervisor, operating system all have the necessary support).

I'm not aware of any system where AVX2 is available at the CPU level, but cannot be made usable with just hypervisor and operating system updates or configuration changes, so this system should really be okay.

If you have code you want me to test or try somewhere I can do so .. if I have answered everything you need, do you want me to close it?

If the x86 Koji builders all have the same hardware, they are fine. glibc runs AVX2 tests during the build if the CPU supports and, and I just checked that the glibc-2.29-18.fc30.x86_64 build ran them (successfully).

More interesting would be image/container builders because I cannot test them directly. That's also the area where we had most problems downstream when we did something similar downstream (albeit on Z, not x86).

That being said, I filed this ticket because someone requested it on the releng ticket. If you agree that everything looks ready on the infrastructure side, feel free to close it.

At the moment, all the hardware we use for builders should have been built post 2015. They are pretty uniform so I believe we can say we are ready for this.