Boards that can run Xeon Gold 6330 Qualification (QWAS) CPU and P200 Optane DIMMs?

[JimD3IL]

After discovering that an X12DAI-N6 will NOT run a QS CPU, I find myself in possession of said CPUs, and need a board that won't turn it's nose up at them. Needs to take Optane DIMMs as well. Any suggestions?

 

[RolloZ170]

discovering that an X12DAI-N6 will NOT run a QS CPU

every motherboard can run QS cpu.
if not(expecting correct installations):
cpu or motherboard defective

QWAS ES2 is NOT a QS processor, but should work with BIOS up to 1.5

 

[RolloZ170]

QS always show full Brand String of CPU in BIOS/CPU-Z e.g. "Intel Xeon 6330" ( "Genuine intel(R) 0000@" is ES)
QWAS is a 16 core, possible a predecessor of Gold 6346
Xeon Gold 6330 is 28 core
QWAQ matches 28 core and is also ES2 worked on X12SPM with BIOS 1.4a but i have no screenshot witth it.

 

[RolloZ170]

and need a board that won't turn it's nose up at them.

sorry but i have to ask:
have you ever installed a LGA4189 processor ?
install of a single cpu in second socket and turn on will kill the X12 motherboard.
you need to turn all 4 screws crosswise alternating (one turn per stroke) until the end, better use a torque screwdriver first time.
down grade to e.g. BIOS 1.4 - possible Supermicro excluded ES from Platform Boot Policy in latest BIOS.

 

[RolloZ170]

QWAQ on X12DAi BIOS 1.1a

 

[JimD3IL]

Rollo- Thanks. I thought it was A QS sample.

▎ To answer your question: yes, I torqued all four nuts crosswise (no torque wrench though- but about 8lbs of torque). And no, I have never powered the board with a single CPU in socket 2 — CPU1 only. I can't speak for the previous owner, who had it briefly.
▎
▎ The reason I think it's the board and not the CPUs: I bought a retail SRKHM Xeon Gold 6330 (28-core, full brand string) and it fails identically. Diagnosis so far, all via BMC/Redfish:
▎
▎ - Port-80 POST code never leaves 0xFF — not a single code is ever written
▎ - BootProgress stuck at 0x01; fans run, power LED solid green
▎ - Microcode never loads: ucode_patch_ver 0x1, versus 0xD0003D1 in the BMC's stale record from the last good POST (2× 6330, 512 GB)
▎ - BMC reads the CPU over PECI at address 0x30 (so it is in CPU1) and gets CPUID 0x606A6, but every uncore register and MSR read fails
▎ - All 46 sensors read "Absent"
▎ - CMOS cleared (battery + JBT1); BIOS 2.4 and BMC 01.07.13 both re-flashed with signed Supermicro images via SAA out-of-band — no change
▎ - Socket inspected after removing the heatsink: no visible bent or missing pins
▎
▎ So the board never executes a single BIOS instruction. It's going back to seller.
▎
▎ Two questions if you have a moment:
▎ 1. Does that signature (POST 0xFF + no microcode + dead sensors) match anything you've seen before?
▎ 2. For the QWAS ES2 pair — is the requirement a board already sitting on BIOS ≤1.5, or can an X12 RoT board be downgraded? I tried 1.1b and it would not apply, which I suspect is RoT rollback protection.

▎ Danke Rollo —
▎ Zu deiner Frage: Ja, ich habe alle vier Muttern über Kreuz angezogen, jeweils eine Umdrehung pro Durchgang, mit Drehmomentschlüssel nach Vorgabe. Und nein, ich habe das Board nie mit einer einzelnen CPU im zweiten Sockel gestartet — ausschließlich CPU1. Für den Vorbesitzer, der es nur kurz hatte, kann ich das allerdings nicht garantieren.
▎
▎ Warum ich das Board und nicht die CPUs verdächtige: Ich habe mir eine Retail-CPU SRKHM Xeon Gold 6330 (28 Kerne, vollständiger Brand String) gekauft — der Fehler ist exakt derselbe. Bisheriger Befund, alles über BMC/Redfish ausgelesen:
▎
▎ - Der Port-80-POST-Code bleibt durchgehend auf 0xFF — es wird nie auch nur ein einziger Code geschrieben
▎ - BootProgress hängt auf 0x01; Lüfter drehen, Power-LED leuchtet konstant grün
▎ - Mikrocode wird nie geladen: ucode_patch_ver 0x1, gegenüber 0xD0003D1 im veralteten BMC-Eintrag vom letzten erfolgreichen POST (2× 6330, 512 GB)
▎ - Der BMC liest die CPU per PECI unter Adresse 0x30 aus (sie sitzt also in CPU1) und bekommt CPUID 0x606A6 — aber sämtliche Uncore-Register- und MSR-Zugriffe schlagen fehl
▎ - Alle 46 Sensoren melden „Absent"
▎ - CMOS gelöscht (Batterie raus + JBT1 gebrückt); BIOS 2.4 und BMC 01.07.13 beide per SAA out-of-band mit signierten Supermicro-Images neu geflasht — keine Änderung
▎ - Sockel nach Abnahme des Kühlers geprüft: keine sichtbar verbogenen oder fehlenden Pins
▎
▎ Das Board führt also nie eine einzige BIOS-Instruktion aus.
▎
▎ Zwei Fragen, falls du kurz Zeit hast:
▎ 1. Sagt dir diese Signatur (POST 0xFF + kein Mikrocode + tote Sensoren) etwas — hattest du so etwas schon einmal?
▎ 2. Zum QWAS-ES2-Paar: Braucht man zwingend ein Board, das bereits auf BIOS ≤1.5 steht, oder lässt sich ein X12-RoT-Board downgraden? Ich habe 1.1b versucht, es wurde nicht übernommen — ich vermute den Rollback-Schutz des RoT.

 

[RolloZ170]

1. Does that signature (POST 0xFF + no microcode + dead sensors) match anything you've seen before?

FF is reset state. means nobody wrote to Port80 IO register.

2. For the QWAS ES2 pair — is the requirement a board already sitting on BIOS ≤1.5, or can an X12 RoT board be downgraded?

normaly allowed to roll back to 1.1 (can be changed if BIOS > 2.0 )

I tried 1.1b and it would not apply, which I suspect is RoT rollback protection.

you need a full BIOS POST to see new version in BMC inventory.
you get a message if downgrade is not allowed.

 

[RolloZ170]

2. For the QWAS ES2 pair — is the requirement a board already sitting on BIOS ≤1.5,

i sold all 4189 boards and processors. that time BIOS 1.5 was latest. don't know if you can use 2.x BIOS.
a year ago i would have sayd "sure" but i was wrong many times.

 

[JimD3IL]

Thanks. This board came with 2.4. I got 1.1b from Supermicro and flashed it but it never got further than staging. If these boards are now unable to be rolled back to 1.5 or earlier when they come with 2.4 flashed, then buyers should be very careful about buying ES or QS chips.

 

[RolloZ170]

This board came with 2.4. I got 1.1b from Supermicro and flashed it but it never got further than staging

as i sayd. if the bIOS is not rejected it is flashed. but you need a POST with a working CPU to see BIOS version in BMC inventory,

 

[RolloZ170]

If these boards are now unable to be rolled back to 1.5 or earlier when they come with 2.4 flashed, then buyers should be very careful about buying ES or QS chips.

as i sayd. QS work on EVERY motherboards. QS is production silicon just ES bit set (no intel support, no warranty after launch)

 

[RolloZ170]

have found a screentshot of mine with X12SPM QWAT and BIOS 1.6 running.
QWAT has same compatibility than QWAS.

 

[JimD3IL]

Rollo,
Thanks. It was the motherboard. I sent it back and got a brand new one from China. I dropped in the QWAS chip (by mistake, but no problem) and it booted with *BIOS 1.9* so yes its good. DO you know if there are any stability problems with these QWAS chips and Linux kernal?

Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 46 bits physical, 57 bits virtual
Byte Order: Little Endian
CPU(s): 32
On-line CPU(s) list: 0-31
Vendor ID: GenuineIntel
Model name: Genuine Intel(R) CPU 0000%@
CPU family: 6
Model: 106
Thread(s) per core: 2
Core(s) per socket: 16
Socket(s): 1
Stepping: 6
CPU(s) scaling MHz: 30%
CPU max MHz: 3600.0000
CPU min MHz: 800.0000
BogoMIPS: 5800.00
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 art 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 3dnowprefetch cpuid_fault epb cat_l3 ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local split_lock_detect wbnoinvd dtherm ida arat pln pts vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid fsrm md_clear pconfig flush_l1d arch_capabilities
Virtualization features:
Virtualization: VT-x
Caches (sum of all):
L1d: 768 KiB (16 instances)
L1i: 512 KiB (16 instances)
L2: 20 MiB (16 instances)
L3: 36 MiB (1 instance)
NUMA:
NUMA node(s): 1
NUMA node0 CPU(s): 0-31
Vulnerabilities:
Gather data sampling: Vulnerable
Ghostwrite: Not affected
Indirect target selection: Mitigation; Aligned branch/return thunks
Itlb multihit: Not affected
L1tf: Not affected
Mds: Not affected
Meltdown: Not affected
Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Old microcode: Not affected
Reg file data sampling: Not affected
Retbleed: Not affected
Spec rstack overflow: Not affected
Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Spectre v2: Mitigation; Enhanced / Automatic IBRS; IBPB conditional; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop
Srbds: Not affected
Tsa: Not affected
Tsx async abort: Not affected
Vmscape: Not affected

 

[RolloZ170]

DO you know if there are any stability problems with these QWAS chips and Linux kernal?

i never had any issues in windows. did not use linux.
e.g. i had issues with a retail cpu with defective core, so not all problems are caused by ES cpu.

 

[RolloZ170]

I dropped in the QWAS chip (by mistake, but no problem) and it booted with *BIOS 1.9* so yes its good.

ca you try with BIOS 2.4 and latest BIOS ?