What controls Turbo Core in Xeons?

Page 89 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

rottwag

Member
Apr 2, 2017
77
11
41
hi everybody!

i need you help!!!

I just changed some minor bios settings yesterday and after that suddently the computer did not post anymore and stopped with error code 68 (visible on the display on the mainboard) !!!

i even usb flashbacked my bios to a stock unhacked version. the bios flashing process seemed to work (pushing bios update button on mainboard, led starts blinking, once it stopps, bios flash was okay and finished)

even after that the computer does not post and show error 68!?!

i did not move the pc or anything! in the www, people with this error checked all components and alle were okay. there is one case, where changing the heat paste brought the success?!

does anybody know this error?

i anyways wanted to plug in my new 2696.... maybe this one runs....
 

Cata40

Member
Mar 2, 2017
156
6
81
i find this with your error
ERROR CODE 68 NO POST. The problem is very annoying. However it has been solved. If you will note it should post in memory slot 1, channel 1. Channel 2 (slot 3) will not post, unless you update your bios. However, updating your BIOS still will not allow channel two memory to be recognized (in bios) and will not be usable (in windows); example: 12 GB installed (4GB usable).

The memory controller is on the Processor of the core I7's. However the problem is not the processor. Even loosening your heatsink as EVGA instructs will not solve the issue. The issue is the backplate for the heatsink. You must loosen it to the point that it's still screwed in however it's movable, allow gravity to hold the heatsink on the cpu and don’t forget to reset your BIOS/CMOS via reset switch or by removing your CMOS battery (very important). After you have done this and there is no pressure on your board (CPU area) go through the process of installing your memory. It should POST with all channels recognizable unless of course you do have some kind of hardware failure. It should be noted though that electronic components are very different then mechanical. Since there are no moving parts, failure is rare.

You can tell that it is a backplate issue also because you will also get a 'New CPU installed' message after taking out the memory module for channel 2 (slot 3), it will post only in channel 1 (slot 1 & 2). You see, the BIOS thinks you installed a new Processor because of the excessive pressure in that area caused by the backplate warping the LGA socket and not allowing the pins for the memory controller to make contact. Loosening the backplate and resetting CMOS fixes it.

Once you FIX and TEST everything with the backplate loose and heatsink unscrewed (gravity contact only), then go ahead and tighten but do not over tighten; although I feel I didn’t tighten it too much either, especially with thumb screws ya know! Anyway, tighten the backplate just enough to make contact with board stopping the backplate from moving around. Then install the heatsink, reset CMOS and install memory and fly away!

NOTE: I've noticed that Noctua backplates have a THICKER pad then Thermalright. The THICKER BACKPLATE PAD allows the backplate to be slightly higher off the back of the LGA socket or CPU area which helps prevent this problem. The Thermalright backplate pad is thinner then Noctua's which will apply more pressure against the back of the LGA socket when tightened. My guess is that you probably do not have this problem if you have Noctua, but I guess variations exist in everything.

I hope this helped, I know it's long, but so is the process of elimination!
 

sciff

Member
Mar 6, 2017
136
52
71
Cata40, since Noctua's backplate is mentioned, the guy is probably talking about socket 1151 not 2011-3 where the backplate is initially installed on the motherboard and provides sufficient hardness to prevent the motherboard from being significantly bent under the weight of CPU cooler heatsink.

What method and ucode do you use to boost the clocks with the ASUS Z10PE-D8 WS and the 2696v3?
I used the taichi x99 to this date but I don't like it. (18 cores at 3.2Ghz)
The ASUS Z10PE-D8 WS is on the way (I only have 1x E5 2696v3 for now) that's why I'm asking and trying to save a lot of time because there are quite a few posts here to read...
Thank you!

https://ibb.co/eUEYZG


xeon.jpg

eUEYZG
I use a modified BIOS with removed µCU (microcode update) and removed ME (Intel Management Engine - the only way to flash such BIOS is to use a SPI Flash device like this one - works flawlessly). For the turbo unlock you just need the µCU to be removed.

You also need to tune the BIOS a little and install an EFI driver. I recommend using one of kjboughton's RC9 drivers (however, I compiled one for myself to get different offsets for my asymmetrical CPU duo - minus 60/50/50 mV for OEM and minus 70/50/50 mV for ES)






How does a [SR1XK] E5-2696v3 (non-ES) would match to a [QGN7] one which is ES or QS but both has the same specs in CPU-Z except the ES mark behind the name in a dual socket asus z10pe d8 motherboard?
They will work together, but in AVX2 software you might observe different behaviour (lower turbo for OEM CPU) if you use 1F microcode which Kris (kjboughton) recommended. I get x27 for the SR1XK and x34 for the QGN7 for some reason. But if you use no microcode at all, you get maximum turbo for both of them in AVX2. With µCU versions 27 and 39 (and 3A for that matter) I did not observe any difference between them, they both got lower AVX2 turbo (x29) which is perfectly normal.
 
Last edited:
  • Like
Reactions: Gellért

sciff

Member
Mar 6, 2017
136
52
71
I have a Z10PE-D16 WS with a single E5-2683 V3 (at the moment, planning on a second once the price go down a little.) Can someone link me to the right bios to use to get the full turbo out of it?

Here

I'd also love to know your bios settings if that would be possible. I have no problem tinkering around with the but if someone has had success fully unlocking it I would love to see what they did to do so an attempt to recreate it.

You can use the settings that I use on my D8 WS mobo, I published the link in the previous post. The difference between two motherboards in terms of BIOS is minimal.

But you also need an EFI driver. This is C612, not X99 ASUS mobo.

P.S. It would be awesome if we can search inside threads...

Actually, you can :)
 

Cata40

Member
Mar 2, 2017
156
6
81
Cata40, since Noctua's backplate is mentioned, the guy is probably talking about socket 1151 not 2011-3 where the backplate is initially installed on the motherboard and provides sufficient hardness to prevent the motherboard from being significantly bent under the weight of CPU cooler heatsink.


I use a modified BIOS with removed µCU (microcode update) and removed ME (Intel Management Engine - the only way to flash such BIOS is to use a SPI Flash device like this one - works flawlessly). For the turbo unlock you just need the µCU to be removed.

You also need to tune the BIOS a little and install an EFI driver. I recommend using one of kjboughton's RC9 drivers (however, I compiled one for myself to get different offsets for my asymmetrical CPU duo - minus 60/50/50 mV for OEM and minus 70/50/50 mV for ES)







They will work together, but in AVX2 software you might observe different behaviour (lower turbo for OEM CPU) if you use 1F microcode which Kris (kjboughton) recommended. I get x27 for the SR1XK and x34 for the QGN7 for some reason. But if you use no microcode at all, you get maximum turbo for both of them in AVX2. With µCU versions 27 and 39 (and 3A for that matter) I did not observe any difference between them, they both got lower AVX2 turbo (x29) which is perfectly normal.
I'm sorry, I'm not so good with English.
Are you the only ones who had the problem with asus board to give this error 68, why did i hear when i'm in the forum
I do not need efi, it works very well without.
I have 424 with 8400 score in bech of cpu-z and my processor can both with or without an efi driver
And in CINEBENCH_R15 have 2797
It goes decent, I have very good temperatures, why would I change?
More can not be done yet
 

rottwag

Member
Apr 2, 2017
77
11
41
Hi,

I tried to solve my starting issues yesterday without success:

As written in other threads already i checked all components, but they are ok.

With only Mainbord, PSU and RAM installed, i also get error 68!

As i cannot believe that my eom xeon broke from one to the other day by just changing unimportant bios settings, i assume that my Mainboard became broken....

I will order a new x99 Mainboard to check if at least one of the xeons boot... What else could i do?

Any recommandations which to by? Asus again? Or better Asrock?

I am sad
Andy
 
Last edited:

Cata40

Member
Mar 2, 2017
156
6
81
I think you have this problem for a long time, just now, at a change, it has come to the surface.
We are so many that changed the normal bios with a modic one and put the eif at the boot, but I have not heard of it
He's looking for google for those bugs, from what it's all about
You can find other cards, just the same chipset, so I'm pretty much the same
I do not think it burned, something is happening
If it burned, you did not have a screen or see those figures down on the board
 

rottwag

Member
Apr 2, 2017
77
11
41
does RAM compatibility depends on the mainboard or on the CPU that is pluged in?
 
Last edited:

txgy

Member
Apr 24, 2017
29
1
41
Same here too. Looks like 1.825volt is a limit, when you go above vcc lock doesn't works anymore.
Hi, MOF
can you make these efi files?
-100mv Vcore, -50 Cache, -50 SA, 1.8V VCCin powercut
-110mv Vcore, -50 Cache, -50 SA, 1.8V VCCin powercut
-120mv Vcore, -50 Cache, -50 SA, 1.8V VCCin powercut
Thank you !
Do not be too tired !
 

GTZ

Member
Jan 25, 2017
36
12
41
Why is replaced by F2 to F4

Like the pic,01 00 00 00 38 00 00 00 16 20 28 03 F2 06 03 00 change to01 00 00 00 38 00 00 00 16 20 28 03 F4 06 03 00


https://pan.baidu.com/s/1o8Ue3yE

Becuase some motherboards will not boot if you delete the microcode. For instance, AsRock will boot fine without a microcode, so simply deleting it thru UBU works. However MSI requires a mircocode to boot. So simply replacing that value will still boot into BIOS and the hack would still work since it senses a BIOS but it is not applied since the string is incorrect.
 

Gellért

Member
May 29, 2017
30
6
41
Cata40, since Noctua's backplate is mentioned, the guy is probably talking about socket 1151 not 2011-3 where the backplate is initially installed on the motherboard and provides sufficient hardness to prevent the motherboard from being significantly bent under the weight of CPU cooler heatsink.


I use a modified BIOS with removed µCU (microcode update) and removed ME (Intel Management Engine - the only way to flash such BIOS is to use a SPI Flash device like this one - works flawlessly). For the turbo unlock you just need the µCU to be removed.

You also need to tune the BIOS a little and install an EFI driver. I recommend using one of kjboughton's RC9 drivers (however, I compiled one for myself to get different offsets for my asymmetrical CPU duo - minus 60/50/50 mV for OEM and minus 70/50/50 mV for ES)







They will work together, but in AVX2 software you might observe different behaviour (lower turbo for OEM CPU) if you use 1F microcode which Kris (kjboughton) recommended. I get x27 for the SR1XK and x34 for the QGN7 for some reason. But if you use no microcode at all, you get maximum turbo for both of them in AVX2. With µCU versions 27 and 39 (and 3A for that matter) I did not observe any difference between them, they both got lower AVX2 turbo (x29) which is perfectly normal.

Thank you very much for the detailed reply and the pictures! Same board same second cpu and same cpu coolers will arrive soon haha. Will try out what you recommended.
 
  • Like
Reactions: sciff

Gellért

Member
May 29, 2017
30
6
41
does RAM compatibility depends on the mainboard or on the CPU that is pluged in?

i ask because i got error 55 on the Asus with my new 2696 ES... my old oem xeon worked properly with the Ram (before it got broken). its Kingston hyper fury x 2400er.

It can be CPU error
Even if it's not fitted properly can cause this problem but the socket has not much play with the cpu and with the cooler on it there's not much chance for this.
Faulty CPU can be
 

sciff

Member
Mar 6, 2017
136
52
71
Thank you very much for the detailed reply and the pictures! Same board same second cpu and same cpu coolers will arrive soon haha. Will try out what you recommended.
Haha! Good for you! I look forward to hearing from you how it goes! :)

Let me just give you a hint on how to install those coolers:

 
Last edited:
  • Like
Reactions: Gellért

Gellért

Member
May 29, 2017
30
6
41
Haha! Good for you! I look forward to hearing from you how it goes! :)

Let me just give you a hint on how to install those coolers:


Thank you very much for the details!
It turned out that I'll get 2 pieces of QGN7. I also have a SR1XK. I guess I keep the 2 QS chips and sell the SR1XK and then I'm not going to use 2 sligtly different ships. (since they not that different like an earlier ES chip as far as I know) What do you think? On the other hand the same principles apply for the turbo unlock with the 2xQS chips, right?
 

imimixa

Junior Member
Jan 23, 2018
18
2
41
Becuase some motherboards will not boot if you delete the microcode. For instance, AsRock will boot fine without a microcode, so simply deleting it thru UBU works. However MSI requires a mircocode to boot. So simply replacing that value will still boot into BIOS and the hack would still work since it senses a BIOS but it is not applied since the string is incorrect.

Thank you.
 

rottwag

Member
Apr 2, 2017
77
11
41
It can be CPU error
Even if it's not fitted properly can cause this problem but the socket has not much play with the cpu and with the cooler on it there's not much chance for this.
Faulty CPU can be

Yea... With my oem CPU i do not get error 55... It stops later in the post process with code 68.

The new Motherboard arrived today, i will plug my ram/cpus later and check if it works and with what CPUs.... Wish me luck...
 
Last edited:

rottwag

Member
Apr 2, 2017
77
11
41
Could anybody name me memory that is known for step1 xeons on asus boards?

Currently installed is 4x 8gb Kingston hyper fury x 2400
 
Last edited:

rottwag

Member
Apr 2, 2017
77
11
41
Update: i get q-code 55 with the g.skill memory as well :-((

PS: ordered a 2696 v3 Qgn7...
 
Last edited:

Gellért

Member
May 29, 2017
30
6
41
Update: i get q-code 55 with the g.skill memory as well :-((

PS: ordered a 2696 v3 Qgn7...

I guess stepping 1 CPUs are pickier with RAMs.
On the other hand I think Xeon CPUs are designed to work with server RAMs (ECC, registered) but they can work in a consumer x99 board with consumer RAM (like you mentioned) anyway..
 

rottwag

Member
Apr 2, 2017
77
11
41
Hi all,

i got my Xeon 2696 V3 QGN7 running on my Asus X99-A :) ....very happy...

Now I need to start from scratch to unlock it BIOS wise....

For the record: This proofs that my OEM Xeon 2618L just died during a computer restart!! I find this really unlikely as this CPU is meant for stable 24/7/365 use - ok on 2.5Ghz instead of 3.5Ghz ... but as i never got too hot and Voltage was even lower than specified.... i really do not understand how this could happen.

What I changed in the BIOS before it died was FIVR settings (I think the one, that disables the connection between board + CPU or something)....

Do you think it could have happend due to "bad" BIOS settings??

Let me know - because , i do NOT want to kill my newly installed 2696 V3 !

Best
Andy
 

sciff

Member
Mar 6, 2017
136
52
71
What I changed in the BIOS before it died was FIVR settings (I think the one, that disables the connection between board + CPU or something)....

Do you think it could have happend due to "bad" BIOS settings??

I detect empirical evidence here that your E5-2618L v3 died because you disabled this feature.
 

rottwag

Member
Apr 2, 2017
77
11
41
You really think this is possible?

i will look up how exactly this bios setting is labeled.

But it was a tip i got from a user here... Seems as if it at least did not harm other CPUs?
 

rottwag

Member
Apr 2, 2017
77
11
41
Hi,

I just realized, that the temperatures read different depending on the tool. e.g. in HWinfo, it seems as if the core temps are approx 10 degrees lower than in other tools!!
This scares me, because I was happy to get 75C running prime.. but if this was actually 85C instead, that would be bad....

Can anybody tell me what "Tjunction" value I need to add to the tool so that i get the correct temperatur?

Besides that. Could I use the Tcase temp`? I know that this value is read from the top of the SKU and is a bit lower - but at least there is an official limit from Intel what is 76C for the 2696 V3....

why is this so complicated.

PS: Have a Noctua D15s + 3 Silent Wings in the Case