Sandy Bridge-E VT-d Virtualization broken w/C1 stepping

Idontcare

Elite Member
Oct 10, 1999
21,118
58
91
It's one thing for "errata" to be discovered some time after a product has been released and then to manufacture a revised processor, but it doesn’t seem right for Intel to release processors with a known major feature fault like this, especially as most buyers are unlikely to know about it and Intel is even less likely to shout about it.

Reminds me of the 23.976 fps issue.

http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/7

Bottom line is that the project triangle is reaffirmed yet again - you can only have 2 of any 3 of the following when getting a product developed and to market:
320px-Project-triangle.svg.png


Unless people have unrealistic expectations of Intel, we have already accepted that Intel is not going to release an erratum-free product at the same cost and within the same timeline.

The techpower-up author seems to have unrealistic expectations. I'm sure that the people who don't care about VT-d in SB-E will be all the happier that its not further delayed but cost the same or more expensive but not delayed.

For the customers who want VT-d, those chips aren't going to come any sooner regardless whether the C1 steppings are sold or withheld from the market, so why withhold them and delay the release?
 

moriz

Member
Mar 11, 2009
196
0
0
for a flagship product, especially a flagship INTEL product, it is entirely reasonable to assume that it will be errata free.

also, one of the key roles of SB-E is for hardware virtualization. without it, it can be argued that they are entirely worthless, since a good portion of its users will be using them to run VMs exclusively.
 

Tanclearas

Senior member
May 10, 2002
345
0
71
The description of what VT-d is exactly is blown a little out of proportion. It's not that ALL virtualization capabilities are not present. In fact, VT-d isn't even an option in some of the original Sandy Bridge models, although they still supported VT-x

http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/2

VT-d is really just starting to be taken advantage of in some of the popular virtualization options. Sure, at the enterprise level there are more virtualization options, but I can tell you that the enterprise level customers are highly unlikely to be the ones jumping immediately to SBe.

Who knows. There might just be a better CPU choice for virtualization in the near future anyway. ;)
 

Tanclearas

Senior member
May 10, 2002
345
0
71
for a flagship product, especially a flagship INTEL product, it is entirely reasonable to assume that it will be errata free.

also, one of the key roles of SB-E is for hardware virtualization. without it, it can be argued that they are entirely worthless, since a good portion of its users will be using them to run VMs exclusively.

It is absolutely not reasonable for Intel to throw a huge amount of inventory in the garbage because a very tiny group of users would be affected by disabling the feature affected by the bug.

Once again, VT-d is only one of the virtualization features of Intel's chips, and not all virtualization programs support VT-d anyway.
 

Rifter

Lifer
Oct 9, 1999
11,522
751
126
for a flagship product, especially a flagship INTEL product, it is entirely reasonable to assume that it will be errata free.

Wow get a grip, if you look over the changelogs for the steppings of any CPU you are going to find they fix numerous(usually over 50-100) bugs with every stepping and revision. There is no such thing as a errata free CPU.
 

moriz

Member
Mar 11, 2009
196
0
0
Wow get a grip, if you look over the changelogs for the steppings of any CPU you are going to find they fix numerous(usually over 50-100) bugs with every stepping and revision. There is no such thing as a errata free CPU.

fine then, let me rephrase: it is entirely reasonable to expect intel to release a flagship product that has all of its features enabled.

happy now?
 

grkM3

Golden Member
Jul 29, 2011
1,407
0
0
I love how sites use this to make it sound like its such a bad thing.I dont even think vtd is supported on regular sandys and is a xeon feature.

and Im sure intel is not even mass producing anything yet so they dont need to throw anything away.
 

Rifter

Lifer
Oct 9, 1999
11,522
751
126
fine then, let me rephrase: it is entirely reasonable to expect intel to release a flagship product that has all of its features enabled.

happy now?

that is a much better statement, and one i will agree with :)
 

Rifter

Lifer
Oct 9, 1999
11,522
751
126
I love how sites use this to make it sound like its such a bad thing.I dont even think vtd is supported on regular sandys and is a xeon feature.

I agree.

While this is dissapointing as long as they figure it out by the Xeon SB-E line i dont see it effecting alot of people.
 

moriz

Member
Mar 11, 2009
196
0
0
I love how sites use this to make it sound like its such a bad thing.I dont even think vtd is supported on regular sandys and is a xeon feature.

and Im sure intel is not even mass producing anything yet so they dont need to throw anything away.

vt-d is supported on some sandy bridge CPUs. curiously enough, it is supported for the 2600, but not the 2600k, same for the 2500 and 2500k.
 

grkM3

Golden Member
Jul 29, 2011
1,407
0
0
Here is a little from another site.


The only good news here for Intel, if you look at that way, is that its current Extreme Edition CPU's don't support VT-d, neither does the Sandy Bridge K-SKU CPU's, so as far as the consumer platform is concerned it looks like there won't be any further delays, but the initial batch should be C1 stepping rather than C2. The C1 stepping should be in production by now, or by the latest by next week, although we don't know when Intel will kick off production of the C2 stepping, but it's very possible it won't be until next year as qualification samples of the C2 stepping isn't expected until the very end of the year......
 

grkM3

Golden Member
Jul 29, 2011
1,407
0
0
vt-d is supported on some sandy bridge CPUs. curiously enough, it is supported for the 2600, but not the 2600k, same for the 2500 and 2500k.

sneaky intel!!!

they lock it out of overclocking power houses to force people that need it to buy higher core count xeons.

If a 2600k had it eneable and were overclocked to 4.7 ghz it would mop the floor with a quad core xeon using vtd

see thats eacltly what I ment about sites that spin this crap,Everyone with a 2600k doesnt even have this feature period and you prolly would never use it or known about it if it wernt for the problem on the new sandy e's

Ill bet the first 6 cores will have it locked out also to keep people paying extra for the xeon version of the chip to use it.
 
Last edited:

moriz

Member
Mar 11, 2009
196
0
0
sneaky intel!!!

they lock it out of overclocking power houses to force people that need it to buy higher core count xeons.

If a 2600k had it eneable and were overclocked to 4.7 ghz it would mop the floor with a quad core xeon using vtd

see thats eacltly what I ment about sites that spin this crap,Everyone with a 2600k doesnt even have this feature period and you prolly would never use it or known about it if it wernt for the problem on the new sandy e's

Ill bet the first 6 cores will have it locked out also to keep people paying extra for the xeon version of the chip to use it.

maybe vt-d can limit the overclocking capability of a chip. after all, it is one extra feature to go wrong. it might also be hard to test for stability in an overclocked chip. last thing we need is a bunch of pseudo-servers with wonky IOMMU support.

but then again, this is intel we are talking about, the king of sometimes dubious product segmentation schemes.
 

imaheadcase

Diamond Member
May 9, 2005
3,850
7
76
Read this in the comments from that site. Makes sense to me, its a non issue.


by Sihastru (October 3rd - 1:55 PM) - Reply

by: qubit

This part of the "news" is your own addition and has a nice scent of bias. You can say that it's your own analysis of the news bit on VR, but there are a few flaws with it.

The most important thing to note is that VT-d is not present on desktop platforms. Not only the CPU has to support VT-d, but also key motherboard components (NB/SB). On desktop consummer products these "features" are disabled. They are also disabled on desktop consummer CPUs.

Enterprise customers are not "most buyers" and they do know exactly what features they need and what they don't need. Also they would not normally buy a consumer grade desktop computer. They will buy enterprise grade workstations and servers, which is the Xeon brand, with Xeon compatible motherboards that will support VT-d if they indeed need this feature.

Games, like "first person shooters", are not an really what servers are for. Your trying to plant an idea in the minds of gullable TPU readers (I really hope they buy Bulldozers, I don't want them on the Intel camp), that SB-E has some major flaw that will affect their gaming performance. In reality this doesn't concern anyone since VT-d is disabled at the hardware level on consumer grade desktops.

So why is a bug in a disabled feature important? It's not.

By the time we get Xeons in the channel, they will all be C2. There will be no need for any recalls and "free replacements".

Sorry to lash out like this, but I thought TPU was above this.
 

IntelEnthusiast

Intel Representative
Feb 10, 2011
582
2
0
While a number of the 2nd generation Intel® Core™ processors (none of the “K” processor or processor with a “5” on the end of their part # like the Intel Core i3-2105) do support VT-d (Virtualization with Directed I/O). The only boards that are supporting this feature in the desktop space are using our Q67 chipset (I know that both Asus and Gigabyte have Q67 board supporting VT-d but I haven’t found anyone else). So without both a board and processor that support VT-d, it will not work. Here is a link for the boards that we have support for VT-d.

Christian Wood
Intel Enthusiast Team
 

mv2devnull

Golden Member
Apr 13, 2010
1,495
143
106
I don't thinkt that the "enterprise vs consumer" is so much about "xeon vs 2600" than it is about "Intel vPro platform vs gaming rigs". Has anyone here even considered buying a Q-series (e.g. Q67) motherboard? They are cheaper than fully flavoured P67/Z68 boards.


Ninjaed by IntelEnthusiast.
 

frostedflakes

Diamond Member
Mar 1, 2005
7,925
1
0
for a flagship product, especially a flagship INTEL product, it is entirely reasonable to assume that it will be errata free.

also, one of the key roles of SB-E is for hardware virtualization. without it, it can be argued that they are entirely worthless, since a good portion of its users will be using them to run VMs exclusively.
All CPUs have errata, I think there's even some place on Intel's site where you can go and look up all the known errata for their CPUs. Many errata don't even bother getting fixed because they're very unlikely to be reproduced during normal usage and won't ever affect 99.99% of users. Wouldn't be surprised that was the case for the VT-d erratum on SB-E, you'd think if it was a major bug it would have been caught before production silicon.

AMD's TLB erratum was kind of the same. It really wasn't a big deal and wasn't a bug that users were likely to ever run in to, it's unfortunate they received so much bad press about it. The problem was blown way out of proportion.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,229
9,990
126
for a flagship product, especially a flagship INTEL product, it is entirely reasonable to assume that it will be errata free.

also, one of the key roles of SB-E is for hardware virtualization. without it, it can be argued that they are entirely worthless, since a good portion of its users will be using them to run VMs exclusively.

This. Intel done F***ed Up.

Edit: I'm also not claiming that the CPU should be completely errata-free, but there shouldn't be any show-stopping bugs that block using any of the CPU's major features.

Would it be right to sell CPUs, that perhaps had the PCI-E lanes crippled to 1.0 speeds? Or if HyperThreading were broken?

Edit: For those claiming that this won't affect gaming, what about places that run a virtualized gaming server, with a VM software with VT-d that allows guests to directly access a video card?

Edit:
There is some good news however. For those wanting to continue using Windows XP, the Waimea Bay platform will be compatible with both 32-bit and 64-bit versions of the operating system. However, Intel won’t be providing Rapid Storage drivers for 32-bit XP, which means that the basic Windows drivers will be handling drive access, which may not be optimal. Unsurprisingly, the upcoming Waimea Bay platform will also be supporting the upcoming Windows 8, when released.
What is this BS? No more XP drivers for X79 platform? MS still supports XP SP3 until 2014. It's a supported platform. Yet Intel doesn't want people to use XP on their hardware any more???
 
Last edited:

bryanW1995

Lifer
May 22, 2007
11,144
32
91
maybe vt-d can limit the overclocking capability of a chip. after all, it is one extra feature to go wrong. it might also be hard to test for stability in an overclocked chip. last thing we need is a bunch of pseudo-servers with wonky IOMMU support.

but then again, this is intel we are talking about, the king of sometimes dubious product segmentation schemes.

Ding ding ding!

Curse you, Hector Ruiz!!!!
 

bryanW1995

Lifer
May 22, 2007
11,144
32
91
Read this in the comments from that site. Makes sense to me, its a non issue.


by Sihastru (October 3rd - 1:55 PM) - Reply

by: qubit

This part of the "news" is your own addition and has a nice scent of bias. You can say that it's your own analysis of the news bit on VR, but there are a few flaws with it.

The most important thing to note is that VT-d is not present on desktop platforms. Not only the CPU has to support VT-d, but also key motherboard components (NB/SB). On desktop consummer products these "features" are disabled. They are also disabled on desktop consummer CPUs.

Enterprise customers are not "most buyers" and they do know exactly what features they need and what they don't need. Also they would not normally buy a consumer grade desktop computer. They will buy enterprise grade workstations and servers, which is the Xeon brand, with Xeon compatible motherboards that will support VT-d if they indeed need this feature.

Games, like "first person shooters", are not an really what servers are for. Your trying to plant an idea in the minds of gullable TPU readers (I really hope they buy Bulldozers, I don't want them on the Intel camp), that SB-E has some major flaw that will affect their gaming performance. In reality this doesn't concern anyone since VT-d is disabled at the hardware level on consumer grade desktops.

So why is a bug in a disabled feature important? It's not.

By the time we get Xeons in the channel, they will all be C2. There will be no need for any recalls and "free replacements".

Sorry to lash out like this, but I thought TPU was above this.

If that statement is true then why do 2600 and 2500 have vt-d enabled? It's clearly an issue for some people, though I agree that it doesn't make sense to further delay the sb-e release for something like this that will only affect a small subset of the potential purchasers.
 

Edrick

Golden Member
Feb 18, 2010
1,939
230
106
What is this BS? No more XP drivers for X79 platform? MS still supports XP SP3 until 2014. It's a supported platform. Yet Intel doesn't want people to use XP on their hardware any more???

Do you actually read what you quote and comment to? No where in that article did it state what you think you have read (minus the RST drivers for 32bit version only).
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
wow u guys really expect an ES to be eratta free?

Really?

Some guys cant even find a good bios for their boards... you know what.. infact some board makers cant finalize anything yet in their boards...

I can tell you one thing tho, any benchmarks u see, will be rough estimates on what the processor can do.
And anything that changes will improve b4 it hits retail.

The Gulftown CPU alone went though i belive 2-3 pre retail releases b4 it hit retail.
A0 step (none made retail.)
B0 Step Some made retail,
B1 Step which hit retail...

These Es's are going to be cream A0's, or B0's even..
Think back when the i7's first hit launch... it was released as a C1/C0 which was replaced quickly by a D0.

You guys also see why i tell people to wait... u really want to get stuck with a c1/c0 again, when possibly a D0 type might come around a few months later?
 
Last edited: