ES Xeon Discussion

[Skud]

@Skud - IMV/most likely either just a typo by reseller or it has defective cores which were disabled as a result. It might be better to contact seller with this question to clarify the situation. Intel clearly describes it as 24c
I was trying communicate with the seller offering cheaper version from your links about a different CPU several months ago and despite some strange English/Chinese translation on their end sometimes could get relevant replies.

I’ve reached out to the seller to ask.

I wonder if this is the result of a market strategy change? Maybe originally the 3345 was supposed to be 20c with 3.2ghz base but was changed at the last minute to 24? Maybe these are those super early units?

I can’t think of any other reason both models would identify as a 3345.

 

[Skud]

allmost all ivy bridge-EP es/qs have 2 cores less than prod.units.

Interesting. I guess this makes sense then if they did the same thing with the w-3000 series QS.

 

[Skud]

you haven't noticed the 20C 3345 is not a ES ?

Yes. It’s listed as a QS. There seems to be a trend where the 20c is listed as QS1/QX7K and the 24 is QS2.

 

[Skud]

allmost all ivy bridge-EP es/qs have 2 cores less than prod.units.

CPU-Z "intel xeon W-3345...." is read out of the processor by cpuid instruction.

Oh gotcha. So then it doesn’t make sense that it would have two fewer cores but still identify as a 3345.

 

[RolloZ170]

I wonder if this is the result of a market strategy change? Maybe originally the 3345 was supposed to be 20c with 3.2ghz base but was changed at the last minute to 24? Maybe these are those super early units?

allmost all ivy bridge-EP es/qs have 2 cores less than prod.units.

I can’t think of any other reason both models would identify as a 3345.

CPU-Z "intel xeon W-3345...." is read out of the processor by cpuid instruction.

Interesting. I guess this makes sense then if they did the same thing with the w-3000 series QS.

you haven't noticed the 20C 3345 is not a ES ?

Yes. It’s listed as a QS. There seems to be a trend where the 20c is listed as QS1/QX7K and the 24 is QS2.

sorry, I should have explained it better, QS is ES too.

CPU-Z does NOT show (ES) it's not QS.

Oh gotcha. So then it doesn’t make sense that it would have two fewer cores but still identify as a 3345.

but it's real. W-3345 20 cores 3.2 Ghz. must be some OEM special unit.
btw: no QS will have the Brand-String on the Heatspreader, you see "Intel Confidential" and the QDF(Q???) there.

 

[Skud]

allmost all ivy bridge-EP es/qs have 2 cores less than prod.units.



CPU-Z "intel xeon W-3345...." is read out of the processor by cpuid instruction.



you haven't noticed the 20C 3345 is not a ES ?



sorry, I should have explained it better, QS is ES too.

CPU-Z does NOT show (ES) it's not QS.


but it's real. W-3345 20 cores 3.2 Ghz. must be some OEM special unit.
btw: no QS will have the Brand-String on the Heatspreader, you see "Intel Confidential" and the QDF(Q???) there.

So the seller is saying that the 20c version is “QS1” and won’t work in supermicro boards. The 24c is “QS2” and is almost the same as retail and will work.

That seems to go against the pictures with the 20c version CPUz not showing it being an ES chip, but the 24c does show ES.

20c is QK7K (no listing on CPU world)
24c is QXZJ (https://www.cpu-world.com/sspec/QX/QXZJ.html)

 

[RolloZ170]

So the seller is saying that the 20c version is “QS1” and won’t work in supermicro boards.

i highly doubt that.
QX7K is WS [ICE LAKE] Processor 3.20 GHz, 30M Cache, 20 Cores.
It's on M0 stepping, this processor wasn't included is the final production configuration.
20c QX7K cpuid 606A6 - Ice Lake-SP M0/M1 D0/D1/D2 (LGA4189)
24c QXZJ cpuid 606A6 - Ice Lake-SP M0/M1 D0/D1/D2 (LGA4189)

 

[Skud]

i highly doubt that.
QX7K is WS [ICE LAKE] Processor 3.20 GHz, 30M Cache, 20 Cores.
It's on M0 stepping, this processor wasn't included is the final production configuration.
20c QX7K cpuid 606A6 - Ice Lake-SP M0/M1 D0/D1/D2 (LGA4189)
24c QXZJ cpuid 606A6 - Ice Lake-SP M0/M1 D0/D1/D2 (LGA4189)

I think I’m going to lean towards the 24 core regardless.

Thanks. I appreciate the help. Out of curiosity where do you get the information to look up the Q spec numbers? I can't find anything on QX7K anywhere.

 

[RolloZ170]

Out of curiosity where do you get the information to look up the Q spec numbers? I can't find anything on QX7K anywhere.

there where some leaks(down) and have some sources i can't disclose, sorry.
edit: if you download the early cpu qvl from gigabyte server boards, you have plenty of QDF informations for free.

 

[RolloZ170]

btw: i have a rare special QS of a planned but never released R4 Xeon W

socket P but recognized for platform R4 LGA2066
there is no W-2265 16 core Xeon W for LGA2066, they share silicon.

 

[Dmitrij_S]

Hi,
Does any of the owners of the platinum 8461v use FEM modeling packages (Comsol, Open-Foam etc. ) ? I looked at platinum 8461v for FEM modeling, it would be interesting to compare it with other processors.

 

[SDletmk]

Can anyone confirm if the Intel Xeon Bronze 3408U QS works on the Supermicro X13SEI-F motherboard or Gigabyte MS03-CE0 ATX? Looking through the thread I saw someone say that the QS should work like a production processor, but I also saw that Gigabyte works better than Supermicro for the ES types. Also, will I be making a mistake when using the Xeon Bronze and/or the X13SEI-F or MS03-CE0 ATX? The system I intend to build with these parts will be running multiple instances of Stable Diffusion with multiple graphic s cards, thus the need for a board with many PCIE lanes.

 

[RolloZ170]

Can anyone confirm if the Intel Xeon Bronze 3408U QS works on the Supermicro X13SEI-F motherboard or Gigabyte MS03-CE0 ATX?

both, have one tested on my MS33-AR0, but a buddy runs earlier ES(not QS) on MS03,MD73.
Q1VK cpuid is 806F7 stepping S2 (MCC single die)
looks like this (the ebay sellers pictures are misleading)

 

[nihonjin]

sorry if here isn't the place... but could you tell me which is today the best/cheap option (between 100-200€) on sale of QS/ES scalables 2nd gen or W-3200s processors?

 

[AlexKaz]

generaly all supermicro UP boards do not run ES with B0/C0 (step 4,5) stepping.
the only known SM board with ES(step 4,5) support is X12DPi-N(T)6
stepping D0 (step 6) runs fine on all boards with latest BIOS so far.
ASRock SC621 and Gigabyte MU72/MD72 do support ES.

RolloZ170, thanks a lot for information about lga4189 and happy New Year!
Can you please say, such v1.0 bioses from Supermicro through drunkencat links are official or unofficial? In theory, are X12DDW-A6 and 5352Y ES can normally run together?
X12DDW-A6 can make downgrade to v1.1 bios through IPMI, I have tested it.

 

[RolloZ170]

Can you please say, such v1.0 bioses from Supermicro through drunkencat links are official or unofficial?

all drunkencat BIOS are official.

are X12DDW-A6 and 5352Y ES can normally run together?

you mean 8352y ES ?

X12DDW-A6 can make downgrade to v1.1 bios through IPMI, I have tested it.

maybe this supports B0/C0 like on X12DPi (up to 1.1b) but not tested.
i doubt that until itis confirmed working.

 

[AlexKaz]

you mean 8352y ES ?

Oh, yes, 8352Y ES QV2F CPUID 606A5, I'm sorry.

maybe this supports B0/C0 like on X12DPi (up to 1.1b) but not tested

I'm using 8352Y ES QV2F with X12DPI, the bios is v1.1b. All ok. But it will be interesting to run the system with X12DDW, because of such board is cheap.

 

[RolloZ170]

I'm using 8352Y ES QV2F with X12DPI, the bios is v1.1b. All ok. But it will be interesting to run the system with X12DDW, because of such board is cheap.

yes, X12DPi-(T)N6 i is a big exception...
i don't know any other, but we can not test all.
i have had X12SPL, works only with unofficial Test/Debug BIOS 1.0 (not 1.0a)
i have X12SPM-LN4F, works only with unofficial Test/Debug BIOS 1.0 (not 1.0a)
but stepping D0 (step 6) ES2 like QWAT/QWAS/QWAP/QWAQ are working like a prod.unit on all boards with latest BIOS.

 

[RolloZ170]

But it will be interesting to run the system with X12DDW, because of such board is cheap

?? how much, where ?

 

[RolloZ170]

X12DDW-A6 can make downgrade to v1.1 bios through IPMI, I have tested it.

this is BIOS_X12DDW-1B58_20210510_1.1_STDsp ME Version: 4.4.4.53
well from the ME version that BIOS is good, but i finaly don't which part blocks step 4/5 from working.