Bdver4 when should Zen2

For KVM VPS or virtualized, the cpu march detected is whatever one, your web host configures the KVM VPS host node for and isn't necessarily the same as the actual cpu. They can for instance configure the lowest common march for a particular family for various reasons i.e. they have a mix of several generations of cpus across all VPS host node servers but have a common KVM OS template, so they need to configure KVM VPS for the lowest/oldest cpu in the family.

That is probably what is going on here. But the other reason is GCC 9 is too old to detect AMD EPYC 2 cpus and may need to wait for GCC 10+ compiler. But GCC 9 is meant to support znver2 AMD Zen 2 CPU Support Merged To GCC 9 (-march=znver2) - Phoronix

You can verify by using a newer GCC compiler to see if it's detected i.e. GCC 9+ via command below

Code (Text):

/opt/rh/devtoolset-9/root/usr/bin/gcc -c -Q -march=native --help=target 2>&1 | grep '\-march'

and output for

Code (Text):

lscpu

and output for

Code (Text):

cat /proc/cpuinfo

Currently Centmin Mod 123.09beta01 and newer PHP routlines will set znver2 when GCC compiler march native check returns znver2 and certain cpu vendor id is detected i.e. AuthenticAMD vendor id is detected + 64bit cpu + GCC compiler march returns znver2

i.e.

Code (Text):

lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                48
On-line CPU(s) list:   0-47
Thread(s) per core:    2
Core(s) per socket:    24
Socket(s):             1
NUMA node(s):          1
Vendor ID:             AuthenticAMD
CPU family:            23
Model:                 49
Model name:            AMD EPYC 7402 24-Core Processor
Stepping:              0
CPU MHz:               1500.000
CPU max MHz:           2800.0000
CPU min MHz:           1500.0000
BogoMIPS:              5599.94
Virtualization:        AMD-V
L1d cache:             32K
L1i cache:             32K
L2 cache:              512K
L3 cache:              16384K
NUMA node0 CPU(s):     0-47
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl xtopology nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme retpoline_amd ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca

 

Another check to do is with command

Code (Text):

/opt/rh/devtoolset-9/root/usr/bin/gcc --target-help | grep -A40 'march=CPU'

does seem znver2 is a valid march for GCC 9 via devtoolset-9 but whether it detects your AMD cpu as such depends on above mentioned factors too

Code (Text):

/opt/rh/devtoolset-9/root/usr/bin/gcc --target-help | grep -A40 'march=CPU' 
  -march=CPU[,+EXTENSION...]
                          generate code for CPU and EXTENSION, CPU is one of:
                           generic32, generic64, i386, i486, i586, i686,
                           pentium, pentiumpro, pentiumii, pentiumiii, pentium4,
                           prescott, nocona, core, core2, corei7, l1om, k1om,
                           iamcu, k6, k6_2, athlon, opteron, k8, amdfam10,
                           bdver1, bdver2, bdver3, bdver4, znver1, znver2,
                           btver1, btver2
                          EXTENSION is combination of:
                           8087, 287, 387, 687, cmov, fxsr, mmx, sse, sse2,
                           sse3, ssse3, sse4.1, sse4.2, sse4, avx, avx2,
                           avx512f, avx512cd, avx512er, avx512pf, avx512dq,
                           avx512bw, avx512vl, vmx, vmfunc, smx, xsave,
                           xsaveopt, xsavec, xsaves, aes, pclmul, fsgsbase,
                           rdrnd, f16c, bmi2, fma, fma4, xop, lwp, movbe, cx16,
                           ept, lzcnt, hle, rtm, invpcid, clflush, nop, syscall,
                           rdtscp, 3dnow, 3dnowa, padlock, svme, sse4a, abm,
                           bmi, tbm, adx, rdseed, prfchw, smap, mpx, sha,
                           clflushopt, prefetchwt1, se1, clwb, avx512ifma,
                           avx512vbmi, avx512_4fmaps, avx512_4vnniw,
                           avx512_vpopcntdq, avx512_vbmi2, avx512_vnni,
                           avx512_bitalg, clzero, mwaitx, ospke, rdpid, ptwrite,
                           ibt, shstk, gfni, vaes, vpclmulqdq, wbnoinvd,
                           pconfig, waitpkg, cldemote, movdiri, movdir64b, no87,
                           no287, no387, no687, nocmov, nofxsr, nommx, nosse,
                           nosse2, nosse3, nossse3, nosse4.1, nosse4.2, nosse4,
                           noavx, noavx2, noavx512f, noavx512cd, noavx512er,
                           noavx512pf, noavx512dq, noavx512bw, noavx512vl,
                           noavx512ifma, noavx512vbmi, noavx512_4fmaps,
                           noavx512_4vnniw, noavx512_vpopcntdq, noavx512_vbmi2,
                           noavx512_vnni, noavx512_bitalg, noibt, noshstk,
                           nomovdiri, nomovdir64b
  -mtune=CPU              optimize for CPU, CPU is one of:
                           generic32, generic64, i8086, i186, i286, i386, i486,
                           i586, i686, pentium, pentiumpro, pentiumii,
                           pentiumiii, pentium4, prescott, nocona, core, core2,
                           corei7, l1om, k1om, iamcu, k6, k6_2, athlon, opteron,
                           k8, amdfam10, bdver1, bdver2, bdver3, bdver4, znver1,
                           znver2, btver1, btver2
  -msse2avx               encode SSE instructions with VEX prefix
  -msse-check=[none|error|warning] (default: warning)

 

My packet.net AMD EPYC 2 Rome 7402 bare metal server reported bdver4 incorrect march target with GCC 8.3 as 8.x didn't support znver2.

But my BuyVM KVM VPS with Ryzen 3900X reports znver2 with devtoolset-9 GCC 9

Code (Text):

/opt/rh/devtoolset-9/root/usr/bin/gcc -c -Q -march=native --help=target 2>&1 | grep '\-march'
  -march=                               znver2

Code (Text):

lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                1
On-line CPU(s) list:   0
Thread(s) per core:    1
Core(s) per socket:    1
Socket(s):             1
NUMA node(s):          1
Vendor ID:             AuthenticAMD
CPU family:            23
Model:                 113
Model name:            AMD Ryzen 9 3900X 12-Core Processor
Stepping:              0
CPU MHz:               3792.872
BogoMIPS:              7585.74
Virtualization:        AMD-V
Hypervisor vendor:     Microsoft
Virtualization type:   full
L1d cache:             64K
L1i cache:             64K
L2 cache:              512K
L3 cache:              16384K
NUMA node0 CPU(s):     0
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 mmxext fxsr_opt pdpe1gb rdtscp lm art rep_good nopl extd_apicid eagerfpu pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core retpoline_amd ssbd ibpb vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 arat npt nrip_save umip

Code (Text):

cat /proc/cpuinfo
processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 23
model           : 113
model name      : AMD Ryzen 9 3900X 12-Core Processor
stepping        : 0
microcode       : 0x1000065
cpu MHz         : 3792.872
cache size      : 512 KB
physical id     : 0
siblings        : 1
core id         : 0
cpu cores       : 1
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
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 mmxext fxsr_opt pdpe1gb rdtscp lm art rep_good nopl extd_apicid eagerfpu pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core retpoline_amd ssbd ibpb vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 arat npt nrip_save umip
bogomips        : 7585.74
TLB size        : 1024 4K pages
clflush size    : 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management:

Code (Text):

/opt/rh/devtoolset-9/root/usr/bin/gcc -c -Q -march=native --help=target | egrep '\-mtune|\-mprefer-vector-width|\-march|\-msse|\-msse2|\-msse2avx|\-msse3|\-mavx|\-msse4|\-mssse3|\-mpclmul'
  -march=                               znver2
  -mavx                                 [enabled]
  -mavx2                                [enabled]
  -mavx256-split-unaligned-load         [disabled]
  -mavx256-split-unaligned-store        [disabled]
  -mavx5124fmaps                        [disabled]
  -mavx5124vnniw                        [disabled]
  -mavx512bitalg                        [disabled]
  -mavx512bw                            [disabled]
  -mavx512cd                            [disabled]
  -mavx512dq                            [disabled]
  -mavx512er                            [disabled]
  -mavx512f                             [disabled]
  -mavx512ifma                          [disabled]
  -mavx512pf                            [disabled]
  -mavx512vbmi                          [disabled]
  -mavx512vbmi2                         [disabled]
  -mavx512vl                            [disabled]
  -mavx512vnni                          [disabled]
  -mavx512vpopcntdq                     [disabled]
  -mpclmul                              [enabled]
  -mprefer-vector-width=                none
  -msse                                 [enabled]
  -msse2                                [enabled]
  -msse2avx                             [disabled]
  -msse3                                [enabled]
  -msse4                                [enabled]
  -msse4.1                              [enabled]
  -msse4.2                              [enabled]
  -msse4a                               [enabled]
  -msse5                     
  -msseregparm                          [disabled]
  -mssse3                               [enabled]
  -mtune-ctrl=               
  -mtune=                               znver2
  Known preferred register vector length (to use with the -mprefer-vector-width= option):
  Known valid arguments for -march= option:
  Known valid arguments for -mtune= option:

So most likely Hetzner's AMD EPYC servers aren't configured to expose znver2 march when GCC 9 checks native march support for znver2

 

Yeah the more cpu models you have in the VPS host node mix, then the more likelihood that the web host configures the lower common denominated cpu model for their KVM configuration. Might contact Linode to see if it's something they can work on i.e. as separate AMD EPYC Linode plan line with exposed znver2 march to KVM guest VPS servers.

 

EPYC-IBPB is for AMD EPYC 1st gen not 2nd gen Ubuntu Manpage: qemu-cpu-models - QEMU / KVM CPU model configuration

not sure if there's something for AMD EPYC Rome ?

 

Sweet that is surprisingly fast action on Hetzner's end eventually Helps they reserved a particular VPS plan to only AMD EPYC Rome

 

excellent news there ! Thanks for keeping us informed