Xeon ES/QS Turbo Bins Database

BenchZowner

Member
Nov 16, 2016
36
5
8
35
Let's get a Turbo Bins database going for all the Xeon ES out there.
By Turbo Bins I mean the maximum clocks ( Turbo Boost ) for all different Active Cores loads.
By having this database going, one can easily see the actual clocks a E.S. Xeon will be running at different active cores. No more wondering if a Xeon ES will turbo boost to 3.3GHz with 4 cores active for gaming, or how fast it'll go with 2 cores active for Photoshop.

Please, everybody with a Xeon ES, take a minute ( or even less than a minute! ) to download AIDA64 install/run it and post a screenshot of the
"CPUID" screen ( be careful to allow all the contents of the "Max Turbo Boost Multipliers" field to show ) from the "Motherboard" submenu category on the left hand side of the program.

see example image:


I will be maintaining a list with all the CPUs added here.

So, what we need is, a screenshot of your CPUID in AIDA64 ( edition doesn't matter, and you don't have to purchase it, the trial is fine ) and also include in your post the stepping /sSpec of your CPU.
 
  • Like
Reactions: rrv

BenchZowner

Member
Nov 16, 2016
36
5
8
35
Starting with my E5-2680 v4 E.S.

CPU Model: E5-2680 v4 E.S.
Stepping: QHV7
Base Frequency (Non-Turbo): 2.20GHz
Turbo Bins:
1 Core active: 26x
2 Cores active: 26x
3 Cores active: 24x
4 to 14 Cores active: 23x

AIDA64 Screenshot:

 

Rand__

Well-Known Member
Mar 6, 2014
4,644
922
113
So here it is again;)

CPU Model: E5-2695 v3 E.S.
Stepping: QFQR
Base Frequency (Non-Turbo): 2.30GHz
Turbo Bins:
1 Core active: 28x
2 Cores active: 27x
3 Cores active: 26x
4 to 14 Cores active: 26x

2017-01-18 09_22_26-AIDA64 Extreme  [ TRIAL VERSION ].png
 
  • Like
Reactions: BenchZowner

ranopile

New Member
Nov 18, 2016
4
2
3
CPU Model: E5-2666 v3 QS
Stepping: QFSC (M0)
Base Frequency (Non-Turbo): 2.9 GHz
Turbo Bins:
1 Core active: 33x
2 Cores active: 33x
3to 10 Cores active: 31x

upload_2017-1-18_22-53-51.png
 
  • Like
Reactions: BenchZowner

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
I don't think this is necessarily needed for QS samples but perhaps for ES1/ES2 samples seeing as how the turbo boosts differ and clock speeds are lower than a QS. However it is an interesting idea but perhaps it should be compiled in something like a spreadsheet like the ES Xeons in the spreadsheet of ES's?
 

BenchZowner

Member
Nov 16, 2016
36
5
8
35
I don't think this is necessarily needed for QS samples but perhaps for ES1/ES2 samples seeing as how the turbo boosts differ and clock speeds are lower than a QS. However it is an interesting idea but perhaps it should be compiled in something like a spreadsheet like the ES Xeons in the spreadsheet of ES's?
That's what I thought.

Although, I'd like to have all QS specs included since there are some that aren't exactly matching the retail chips, and anyway, it doesn't hurt to have a "verified" db ;)
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
That's what I thought.

Although, I'd like to have all QS specs included since there are some that aren't exactly matching the retail chips, and anyway, it doesn't hurt to have a "verified" db ;)
Which QS didn't match a retail or production processor? 99% I've seen (that are an actual QS and not called a QS because they are an ES1/ES2) have identical turbo bins and clock speeds as a production processor.
 

BenchZowner

Member
Nov 16, 2016
36
5
8
35
It was a preview sample I had been given a couple of years ago.
Could've been pre-qs.

Anyway, since there's no database with the per core clock speeds available ( unless you have one and I haven't noticed it online ), what's bad according to your opinion to include the QS sSpecs and their corresponding per core turbo bins ?
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
It was a preview sample I had been given a couple of years ago.
Could've been pre-qs.

Anyway, since there's no database with the per core clock speeds available ( unless you have one and I haven't noticed it online ), what's bad according to your opinion to include the QS sSpecs and their corresponding per core turbo bins ?
QS information is already available in CPU-World and the respective turbo bins of an actual production sample are available on Wikipedia's Xeon list which are applicable to a QS since they are practically identical aside from microcode and string. Here's an example:

List of Intel Xeon microprocessors - Wikipedia

I'm not necessarily against the idea of QS turbo bins but they should identical to their respective production parts. If there are QS's that aren't respective of their production parts then we should mention or catalog them. I think getting an ES1/ES2 turbo-bin database is a higher priority than for QS's.
 
Last edited:

BenchZowner

Member
Nov 16, 2016
36
5
8
35
QS information is already available in CPU-World and the respective turbo bins of an actual production sample are available on Wikipedia's Xeon list which are applicable to a QS since they are practically identical aside from microcode and string. Here's an example:

List of Intel Xeon microprocessors - Wikipedia
Not all CPU models are listed with the turbo bins.
And, would it hurt to have every model enlisted regardless ?

I can't help but wonder, arguing ( for the sake or arguing ) over the inclusion of somethings is your passion or contributing ?
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
Not all CPU models are listed with the turbo bins.
And, would it hurt to have every model enlisted regardless ?

I can't help but wonder, arguing ( for the sake or arguing ) over the inclusion of somethings is your passion or contributing ?
If something is missing we can always include it but the focus should be on ES1/ES2 processors. I don't see why there is confusion on this. There's also this list that can be referenced but it doesn't include turbo bins and has some missing information but it's largely complete (compiled by Nathan_P from HardForum):

E5 Xeon list
 
Last edited:

BenchZowner

Member
Nov 16, 2016
36
5
8
35
*specs and values from info I saw on the web*
Which could very easily include errors ( made from OP user ).

On another note, a complete list is simply... better ?
Imagine crawling through taobao, having the list on the second monitor / split-screen and being able to see the individual clocks for each CPU and activity/application at once without having to scroll through wikipedia lists with incomplete entries.

What if it's one or two chips at a very decent price where 1 second can be enough for somebody else to snatch them before you figured out if they suit your needs or not.

p.s. it also serve as an ES1/ES2/QS sSpec match list which will most certainly help avoid ES sold as QS on the web
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
p.s. it also serve as an ES1/ES2/QS sSpec match list which will most certainly help avoid ES sold as QS on the web
I agree with this. Let me make a contribution here.

E5-2670 V3 ES2 (QEYK):

Intel Xeon E5 2670 V3 ES QEYK 12Core 2.2GHz 30M 24T 120W LGA2011-3 Processor CPU | eBay

Clock Speed: 2.2Ghz (versus 2.3Ghz of retail)
Turbo boost: 2.5Ghz (all cores active), 2.8Ghz (one/two cores active versus 3.1Ghz of production)

Let me know how it should be formatted or entered in for the database. I don't have this same processor so I can't say all the turbo bins but it should be similar core count active like the production 2670 V3 for the turbo bins. Retail/OEM E5 2670 V3 has the following turbo bins:

3/3/3/3/3/3/3/4/5/6/8/8

We can therefore conclude that the 2670 V3 ES2 QEYK has the following:

2/2/2/2/2/2/2/?/?/?/5/5
 

BenchZowner

Member
Nov 16, 2016
36
5
8
35
The three questionmarks is the main issue here.
We can't tell what it is unless we have physical access to a QEYK sample.
Could be 4/3/2
but it also could be 2/2/2
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
The three questionmarks is the main issue here.
We can't tell what it is unless we have physical access to a QEYK sample.
Could be 4/3/2
but it also could be 2/2/2
Yea in some cases unless you have AIDA64 with access to one of these ES2's we will be missing some turbo bin information. In these kind of situations I think we should still include these ES2 samples but mention their all core and single/two active core turbo bins which in this case is 2.5Ghz and 2.8Ghz. As I understand it's usually rare for an ES2 to have finalized clock speeds or turbo bins (but there are exceptions like the E5 2630 V4 QK3G).

Also forgot to mention that ES2 is an L0 stepping. Production stepping is M0. In addition to S-SPEC the stepping should be included (as identified by HwInfo64 and not necessarily CPU-Z).
 

BenchZowner

Member
Nov 16, 2016
36
5
8
35
Missing info is not my thing, but it can be accepted :p

Key point is all the entered data to be verified one way or another.

Stepping to be added ;)
 

Nanotech

Active Member
Aug 1, 2016
595
99
28
40
So here's an interesting example as I mentioned before and another addition. This is the ES2 QK3G of the E5-2630 V4 but it has a production stepping of R0 just like a QS even though it's an ES2.

Intel Xeon E5 2630 V4 ES QK3G 2.2GHz 10Core 25MB 85W LGA2011-3 Processor CPU

It's identical clock speed and turbo boost information make it an exception for ES2 samples usually. So the clock speed and the turbo boost information is identical to a QS/production aside from the string of how it's detected:

2/2/2/3/4/5/6/7/9/9

So max clock speed is 3.1Ghz with two cores and 2.4Ghz with 10-8 cores active which is identical to a QS and production processor. The difference is that QKET/QKRH are the QS samples which are identified as an 2630 V4 string (albeit with an ES) and this is identified as a generic ES.