SB-E C2 stepping on Jan. 20th (Rumor)

[Edrick]

http://www.techpowerup.com/156625/I...-Transitioning-to-C2-stepping-in-January.html

The CPUs moving to the C2 stepping are the hexa-core Core i7-3960X (3.3 GHz) and Core i7-3930K (3.2 GHz). Beside the fixed VT-d, the C2 chips will feature new S-spec and MM numbers so a BIOS update for current motherboards will likely be required.

I wonder if these will have better OCing results as well?

Update:
http://www.dvhardware.net/article52906.html

 

[jsedlak]

Interesting... What was wrong with VT-d?

 

[mrjoltcola]

Darn, how do I tell which stepping I have? I bought first wave, on the release date.

CPU-Z shows "Stepping: 6" for my 3930K

And I actually use VMware a lot, just haven't configured VT-d on this station yet. I assume I have a warranty claim now if I was not aware of this?

 

[notty22]

Intel Core i7-3960X and i7-3930K CPUs Transitioning to C2 stepping in January

you must have the earlier stepping.

Sandy Bridge-E VT-d Broken In C1 Stepping, Fixed In C2 Stepping, Shortly After Launch

 

[mrjoltcola]

Anyone have links to any original Intel announcements on this bug? All I can find are the Product Change Notifications that don't detail the changes, but I see no official statements. I called Intel Support, and the first level support tech wasn't aware of the issue, then claimed it would be fixed in BIOS updates to motherboards; when I asked how they would fix a chip bug with BIOS fixes, he offered to escalate the ticket, and I accepted, and am waiting to hear back from level 2.

 

[notty22]

nv mind

 

[Vesku]

http://ark.intel.com/products/63697/Intel-Core-i7-3930K-Processor-%2812M-Cache-3_20-GHz)

They still have VT-d as "Yes", not cool if they need a new revision to fix it.

 

[AdamK47]

Anyone have links to any original Intel announcements on this bug? All I can find are the Product Change Notifications that don't detail the changes, but I see no official statements. I called Intel Support, and the first level support tech wasn't aware of the issue, then claimed it would be fixed in BIOS updates to motherboards; when I asked how they would fix a chip bug with BIOS fixes, he offered to escalate the ticket, and I accepted, and am waiting to hear back from level 2.

I know it's been mentioned on this forum before the CPUs were released. It doesn't affect me since I never planned on running it anyway.

 

[Idontcare]

http://ark.intel.com/products/63697/Intel-Core-i7-3930K-Processor-(12M-Cache-3_20-GHz)

They still have VT-d as "Yes", not cool if they need a new revision to fix it.

To me that would signify that they know the volume of bugged SB-E's out in the wild was going to be small enough that they felt it was worth the expense of replacing them on a customer-by-customer basis when those customers called in as mrjoltcola did, versus listing the initial steppings as lacking while listing later ones as having expanded feature list.

In other words I see it as a calculated move by Intel. How many enthusiasts are going to buy initial stepping SB-E's but actually want to put VT-d to work like mrjoltcola versus adamk47?

Its definitely "not cool" that they are potentially baiting customers into buying chips that claim to have features when they knowingly don't (that's class action liability right there) even if the plan all along was to switch them out with fixed steppings.

Customer's lose time and money with that plan, if that is their plan. How many mrjoltcola's are out there who aren't exactly building these rigs themselves such that they can be expected to have the expertise to pull their SB-E from the socket and properly reinstall another one? That can get expensive once you factor in customer support costs for the OEMs and resellers.

Would be great if our resident Intel Rep (IntelEnthusiast) could chime in here to let us all know what the skinny is with these initial stepping SB-E's and VT-d functionality.

 

[mrjoltcola]

Would be great if our resident Intel Rep (IntelEnthusiast) could chime in here to let us all know what the skinny is with these initial stepping SB-E's and VT-d functionality.

Agreed. The level one support informed me it was a BIOS issue, directed me to contact ASUS, then when I persisted, he asked me to reply to the support ticket with links to where I read that it was a chip bug. I obliged, but I lack confidence that my case is being handled properly if support doesn't even acknowledge the chip flaw.

 

[Grooveriding]

A portion of Intel's response to my inquiry.

'[FONT=&quot]I understand you want to know if Intel will be replacing your Intel® Core™ i7-3930K Processor due to the issue found affecting the VT-D feature.[/FONT] [FONT=&quot] [/FONT]
[FONT=&quot]Intel discovered the Errata BS90 affecting the VT-D feature of these processors but it was fixed already. You can read the [/FONT][FONT=&quot]“Intel® Core™ i7 Processor Family for the LGA-2011 Socket Specification Update PDF” at:[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]http://www.intel.com/content/dam/doc/specification-update/core-i7-lga-2011-specification-update.pdf[/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]In page 13, it shows the “[/FONT][FONT=&quot]Errata summary table” and as the status says it is fixed. In page 41 shows details about it and a recommended work around.[/FONT][FONT=&quot][/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]All the motherboard manufactures should have developed already a new BIOS update to adapt their boards to the fix applied by Intel into the processors. [/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]Replacing the processor won’t resolve the problem if the motherboard manufacturers don’t provide their users with a newer BIOS version containing the new micro-code to support the C-2 step.[/FONT]

[FONT=&quot] [/FONT]
[FONT=&quot]You should be able to use the VT-D with your actual processor with no problems as long as the BIOS of your motherboard contains a new update with the fix. [/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]Either keeping the processor with C-1 step or getting a new one with C-2 step, at the end it will be required the BIOS update. [/FONT]
[FONT=&quot] [/FONT]
[FONT=&quot]The recommendation is to contact ASUS and request them for a BIOS update that involves the fix applied.[/FONT]'


I find the answer erroneous, unless the reason for the upcoming C2 stepping is something other than correcting this issue.


I didn't take the time to read the information in the link as yet, my quick impression is that there is a BIOS work around for the current stepping, but it is not being implemented via the actual CPU vs the upcoming C2 stepping which enables it within the CPU.


Or do I have this all wrong ?

 

[sm625]

If I bought one of these thigns, I'd trade it in for a new stepping. Especially if all it took was a bogus excuse. I wish I could have done that with my Q6600, I would have loved to get a G0...

 

[mrjoltcola]

[FONT=&quot]Intel discovered the Errata BS90 affecting the VT-D feature of these processors but it was fixed already.[/FONT]

...

I find the answer erroneous, unless the reason for the upcoming C2 stepping is something other than correcting this issue.

I feel the same. For some reason in this day and age, people cannot answer a simple question with a simple answer. It takes 10 minutes of run around jargon and a long email and I still don't have an answer.

If I told my software customers who reported a bug that the "bug was fixed already" and left it at that, they would still want to know which version of the code that I fixed it in, whether it affects them, and whether they need to update.

As I read this, our processor does indeed have a bug, but there is some sort of BIOS workaround. This tells me that the feature probably doesn't work in hardware, so there is a software fix for it; unacceptable when the feature in question is hardware virtualization and acceleration.

 

[AdamK47]

After reading their reply, I'm inclined to believe the "BIOS workaround" requires installing the C2 stepping CPU with the newer BIOS applied.

All the motherboard manufactures should have developed already a new BIOS update to adapt their boards to the fix applied by Intel into the processors

That's some clever wordsmithing on Intels part.

Save face points = +10

 

[MrTransistorm]

I can understand the frustration of anyone who would actually use VT-d, but I don't see how anyone can be surprised about this now. News of the new stepping to fix VT-d was widespread at the beginning of October. A few links:

http://www.techpowerup.com/152978/S...ixed-In-C2-Stepping-Shortly-After-Launch.html

http://vr-zone.com/articles/sandy-bridge-e-should-hit-c2-stepping-after-launch/13677.html

http://www.guru3d.com/news/intel-sandy-bridgee-to-get-c2-stepping-after-launch/

Perhaps Intel should have been more vocal about the flaw, but then a lot of people who have no use for VT-d (or even virtualization in general) would have been in an uproar. OMG my CPU is broken! RAEG!

 

[mrjoltcola]

The problem is, none of these links are Intel announcements, and Intel is actually telling us now that there is no processor bug, and that it is a BIOS microcode problem that has been fixed. I have two emails from 2nd level Intel support specifically saying there is no processor with a bug. I really don't know what to think but I'm inclined to believe Intel support before sites like Guru3d anyway. This age of speaking without making a statement is just maddening.

Noone is surprised, I saw an article way back in October, but it was well before release and I trusted Intel to sell me a bug-free product, as do other customers who don't read these articles.

 

[Grooveriding]

It will all come out in the wash eventually.

When they do release the new stepping, comparisons and analysis will be made on the differences and if it contains the fix we'll have our answer.

 

[mrjoltcola]

@Grooveriding: Yes, probably correct. I escalated my ticket again, and will do so until someone gives me a actual explanation of the problem. They are still using customer speak with me at this point.

 

[SexyK]

I wonder if this new stepping on the horizon has anything to do with the fact that I can't seem to find 3930's in stock much of anywhere these days....

 

[Edrick]

I wonder if this new stepping on the horizon has anything to do with the fact that I can't seem to find 3930's in stock much of anywhere these days....

I read that Intel only released 10K of each CPU of the C1 stepping because they knew C2 was just months away.

 

[mrjoltcola]

So help me parse this email from Intel level 2 support.

The workaround on the situation with the VT-d on the processor has been fixed through BIOS update, since it is something related with the micro-code for the C-1 stepping of the processor. As part of business continuity Intel is providing the market with the C-2 stepping for this processor which includes the complete fix for it to interact any BIOS release as of today.

Been this the solution and before we set up any escalation we require the answer of the following questions:

1- How the VT-d situation affected you business wise?
2- How many systems have been affected due to this situation?
3- What specifically do you need from Intel besides the fix mentioned above?

We will be expecting your comments in order to address them and route them accordingly

Sincerely,
<name removed>
Intel Technical Support

I bolded the items that lead me to believe the BIOS fix is indeed a software patch for a bug that needs to be fixed in hardware. Besides that, there is no BIOS update for my board that I see addresses the problem.

 

[Grooveriding]

So help me parse this email from Intel level 2 support.



I bolded the items that lead me to believe the BIOS fix is indeed a software patch for a bug that needs to be fixed in hardware. Besides that, there is no BIOS update for my board that I see addresses the problem.

The last email I received mentioned the C-2 chips will be available after January 20th.

Reading their reply to you, it looks as if they want you to give them an example of how the bug is limiting your use of the processor. On one hand that seems reasonable, on the other, an advertised feature whether you want to use it or not should be there if advertised.

Bad analogy incoming. My case has front USB 3.0 ports and an adapter wire to plug it into a USB 3.0 header on my MB. I don't use it because my motherboard does not support it and I have no use for it because I have no USB 3.0 devices, but it was advertised as being included and is not in my box. Do I ask for a replacement because it was advertised and I deserve to receive all that was said to be included, or forget about it because I don't plan to make use of it ? In a situation like this I feel it is at my discretion and grace not the manufacturers, their position is to service me depending on my wants.

They may be trying to make it difficult and walk a fine line to avoid setting a precedent of allowing RMAs on the release 3930K/3960X chips and finding themselves inundated with RMA requests.

I may push for an RMA myself if there is more clarity shed on the obtuse responses Intel are giving that points to there being a hardware flaw. Not because I need the feature, but because I paid for it and I may want it for some reason down the line.

 

[RavenSEAL]

Darn, how do I tell which stepping I have? I bought first wave, on the release date.

CPU-Z shows "Stepping: 6" for my 3930K

And I actually use VMware a lot, just haven't configured VT-d on this station yet. I assume I have a warranty claim now if I was not aware of this?

SB-E C2 stepping on Jan. 20th

 

[Vesku]

Is it really a big deal for them to have said "no VT-d on this first stepping" and listing it as such in their ARK database?

The Intel response is unacceptable to me, I would respond with "is VT-d working 100&#37; on the hardware level with this C1 stepping, assuming I have the correctly updated BIOS?" Judging from them asking for examples of it impacting your business, this all comes across as quite sketchy behavior from a top tier manufacturer.

 

[Grooveriding]

Wanted to post up my recent correspondence results with Intel in relation to obtaining an RMA from a C-1 stepping to a C-2 stepping SB-E due to the Vt-d issue.

I pursued this issue because I was curious as to how they would handle it and it irks me to see something advertised as an included feature which in actuality was not included.

Some interesting developments.

For a few email exchanges there were continued references to a BIOS update that enables support on 'chips that support vt-d', but as per my previous example of a reply I received from Intel the language being used was very cryptic and lacked anything concrete. It was literally impossible to get a reply specifically saying Vt-d support is enabled by this BIOS update for a C1 stepping SB-E on the CPU hardware its self.

Fortunately this pdf came to my attention which explicitly outlines the reasoning for the change to the C-2 stepping and one of the two given reasons is to correct the Vt-d errata:

http://qdms.intel.com/dm/d.aspx/F051E059-17FA-4B21-92E7-025550FBF529/PCN111178-00.pdf

Here is the interesting part. Intel has updated the Ark database to now reflect that the 3930K & 3960X do not support Vt-d. Previously it did show support and previously meaning when these chips were on the market and available for sale. I'm not implying they are up to anything nefarious but it was definitely irresponsible and makes them beholden to provide RMAs if requested in my book.

Further I was told by Intel that the website shows there is no Vt-D support for the current SB-E processors and never has. Fortunately I had a cached copy of the original page's contents to counter with.

I did not receive a direct response to providing this website history, but did receive notice I can RMA my chip with some caveats, ones that make my stated intent of RMA ineffectual.

I've now been told that I can have an RMA provided but there is no guarantee that I will receive a C-2 stepping as the RMA department will deplete whatever stock they have on hand; C-1 step, without being able to give a specific stepping after the January 20th release of the new stepping.

I expressed my thanks for being given the option of RMA and advised I would be in contact when C-2 stepping chips are on the market to specifically obtain an RMA for said stepping of the chip.

Will update in a few weeks on the outcome.

Edit: Here is a link from Tom's on the matter http://www.tomshardware.com/news/intel-cpu-sandy-bridge-e-c2-virtualization,14252.html

While C1 CPUs are only running the software-accelerated only mode, the C2 now properly supports hardware acceleration on a hosted OS, Intel said. According to the document sent out on December 9, samples of the C2 Sandy Bridge-E processors became available last week. Commercial retail and tray units will begin shipping on January 20.