Intel released critical microcode patch for all i7 980x,Xeon 36../56.. systems

[duploxxx]

just FYI, there are already few with new 6cores

SYSTEM BIOS UPGRADE REQUIRED; Certain Intel Processors May Cause Unpredictable System Behavior
NOTICE: The information in this document, including products and software versions, is current as of the Release Date. This document is subject to change without notice.
Release Date: 2010-04-12

Last Updated: 2010-04-12


--------------------------------------------------------------------------------

DESCRIPTION
CAUTION: This system BIOS upgrade is considered a critical fix and is required to correct the issue detailed below. OEM strongly recommends immediate application of required critical fixes. Failure to perform the required action could leave the workstation or desktop in an unstable condition, which could potentially result in system lockups or failures, or data loss.

Intel has provided a microcode update to OEM as a critical software fix. This Intel microcode update addresses an issue that may result in unpredictable system behavior. OEM is making available system BIOS upgrades that integrate the updated Intel microcode to address this issue. While the potential for occurrence of this issue is extremely rare, OEM strongly recommends upgrading the system BIOS to the version listed in the RESOLUTION section (or later). Intel has indicated that the microcode update has no known measurable performance impact and provides a complete and comprehensive resolution to this issue.

This issue was identified in a synthetic test environment during rigorous quality control testing. To date, OEM has not observed this issue in a production environment.

This issue will be documented in an updated version of the Intel Processor Specification Update for all affected processors.

Click here for the Xeon 3600 Intel Processor Specification Update.

Click here for the Xeon 5600 Sequence Intel Processor Specification Update

Click here for the Intel Core i7 Processor Extreme Edition Specification Update.


IMPORTANT : This condition is not specific to OEM Desktops and can affect any desktop or workstation that uses the Intel Core i7-980X Extreme Edition Processor, Intel Xeon Processor 3600 and 5600 sequences.

While the implications of the issue are difficult to quantify, any of the following symptoms can occur:

The system may stop responding to keyboard or mouse input.
A system operating in a Microsoft Windows environment may generate a blue screen.
A system operating in a Linux environment may generate a kernel panic.

SCOPE
This issue may affect any of the following workstations or desktops that may be offered with these Intel processors; the Intel Core i7-980X Extreme Edition Processor, Intel Xeon 3600 and Intel Xeon 5600 Sequences and the BIOS versions listed below.

To prevent the unlikely occurrence of this issue, upgrade the workstation or desktop system BIOS to a version listed for your specific system. System BIOS listed includes the required Intel microcode updates to address this issue. Applying the latest system BIOS ensures that the most current version of Intel microcode is installed in the system.

[necronomicon]

Source?

[JumpingJack]

The source is Intel's PCNs most likely, this is not atypical as rare errata are discovered companies will issue BIOS updates in the microcode that are then routinely distributed to OEMs and ODMs. AMD does this as does Intel, this is really just someone trying to dish up dirt and blow it out of proportion.

[highoctane]

this is really just someone trying to dish up dirt and blow it out of proportion.

This would seemly fit the reality of forums today, like a big soap opera.

[JumpingJack]

This would seemly fit the reality of forums today, like a big soap opera.

I said the same thing about the TLB errata on Phenom I ... that one was the most blown out of proportion errata I have seen, ever. That one will be referenced historically for years to come as the most over hyped nonsense in the community.

[Clairvoyant129]

The source is Intel's PCNs most likely, this is not atypical as rare errata are discovered companies will issue BIOS updates in the microcode that are then routinely distributed to OEMs and ODMs. AMD does this as does Intel, this is really just someone trying to dish up dirt and blow it out of proportion.

Well seeing as who created this thread, I'm not surprised.

[570091D]

I said the same thing about the TLB errata on Phenom I ... that one was the most blown out of proportion errata I have seen, ever. That one will be referenced historically for years to come as the most over hyped nonsense in the community.

QFT

i remember AMD posting a $100,000 reward for anyone who could replicate the TLB problem in the real world

but hey, at least intel is on the ball and issuing a fix

[saaya]

SYSTEM BIOS UPGRADE REQUIRED; Certain Intel Processors May Cause Unpredictable System Behavior

unpredictable system behaviour? wtf is that supposed to mean?

ah well...
*starts bios update*
"im affraid i cant let you do that saaya..."
huh? what? whos there?

[JumpingJack]

QFT

i remember AMD posting a $100,000 reward for anyone who could replicate the TLB problem in the real world

but hey, at least intel is on the ball and issuing a fix

I have always held a completely neutral position here -- both Intel and AMD are really good on being on the ball, frankly. All processors have errata, most of them are discovered in the last stages of testing, many (a significant portion) appear after the fact.

The point is, both AMD and Intel subject new products to huge amounts of testing spanning most usage cases and almost all of the known commercial software and OSes, these errata are almost always benign or very rare occurence, many don't even get fixed, of those that don't the probability of triggering one is next to nothing.

The only real errata I can recall that caused significant damage, not necessarily to the compute world but more to reputation was the FDIV bug. That was nasty and Intel did not handle it well at all, they learned their lesson there no doubt.

[JF-AMD]

I have to agree 100% that errata happens and we shouldn't be calling it out.

However, I do also agree that if someone is going to market their product as being "more secure" and denigrate their competitor as being "less secure", they open themselves up on things like this.

[kl0012]

I have to agree 100% that errata happens and we shouldn't be calling it out.

However, I do also agree that if someone is going to market their product as being "more secure" and denigrate their competitor as being "less secure", they open themselves up on things like this.

Sh... errata hapens... but what this "more secure" thing has to do with an errata? As I know "more secure" was related to a new "AES" instruction set and as I know Intel didn't labeled competitors as "less secure".

[Hornet331]

I have to agree 100% that errata happens and we shouldn't be calling it out.

However, I do also agree that if someone is going to market their product as being "more secure" and denigrate their competitor as being "less secure", they open themselves up on things like this.

*insert ohyou.jpg in here*

marketing people on both sides brag with stupid things, that are laughable in the eyes of every enthusiast...

[JF-AMD]

Sh... errata hapens... but what this "more secure" thing has to do with an errata? As I know "more secure" was related to a new "AES" instruction set and as I know Intel didn't labeled competitors as "less secure".

http://www.intel.com/pressroom/archi...316comp_sm.htm

"Intel Launches its Most Secure Data Center Processor"

So, if you are claiming that you are most secure, by implication the others are less secure, right?

Most of the threads where the Intel guys were pushing this have been locked/deleted because they became regular troll fests.

One intel person in particular said you should never run virtualization without AES. The funny thing is, any x86 processor can execute AES, and the majority of his company's processors don't have AES instructions built in, only westmere.

There is just a point where taking a position becomes more of liability than a benefit. Security is not the thing I would want to hang my hat on because everyone knows that there will be errata and issues that you need to resolve. A critical microcode patch makes the "most secure data center processor" tag a little harder to believe.

There are a lot of things in their processors that you can talk about that are good, I have never been a big fan of hanging my hat on security because a.) there are always errata (processors are complex) and b.) so much of security is outside of the processor's control.

[josh1980]

JF-AMD-

You are an AMD employee based on your signature. I do not believe it is correct to automatically take the side of someone claiming "Intel Launches its Most Secure Data Center Processor". I take that to mean that Intel is saying it has released THEIR most secure processor. They make no claim whatsoever to how any competitor(s) products function compared to their own.

On the other hand, you have taken the side that Intel has claimed their competitor(s) processors are inferior.

Based on your misunderstanding of that one quote, I have to question your other comments discrediting what you call an "intel person" claiming that "you should never run virtualization without AES". I also question all of these threads that are allegedly locked or deleted. I will, however, be gracious enough to give you the benefit of the doubt that there are deleted threads out there(hence you could never actually defend your position) even if you will not do the same for Intel.

I remember playing Everquest and a few other games that released game updates, and for a short time the game was UNPLAYABLE on the entire line of current AMD processors at the time. At that time, I had always had an Intel computer, and I had no idea where AMD and Intel stood. But I got a laugh out of reading that at the EQ website and I still have the screenshot of the message somewhere. It's fun to see the fanboys be thrown into a position they have zero chance of defending and watch them squirm. (Typically, I've always seen MANY more AMD fanboys than Intel fanboys. Not sure if this is how it is or just luck on where I go).

[kl0012]

There is just a point where taking a position becomes more of liability than a benefit. Security is not the thing I would want to hang my hat on because everyone knows that there will be errata and issues that you need to resolve. A critical microcode patch makes the "most secure data center processor" tag a little harder to believe.

There are a lot of things in their processors that you can talk about that are good, I have never been a big fan of hanging my hat on security because a.) there are always errata (processors are complex) and b.) so much of security is outside of the processor's control.

You probably don't like it but a fact is a fact - CPU which has AES instruction and can handle security algorithms 6-8 times faster is better suited for some type of workloads. Now it seems that you're mixing up a different things. Not every errata causes for security breach. And while I have no hard statistic, I still have a filling that things like microcode update happens not so rarely by both firms. Just this time someone choose to disclose internal information intended to OEMs.

[radaja]

"Intel Launches it's Most Secure Data Center Processor"

come on JF-AMD,this is ridiculous,your trying to make something out of nothing.i agree with josh1980
the quote above clearly say IT'S most secure,meaning intel's. not THE most secure in the whole industry
and AMD is less secure.

[JF-AMD]

Point taken on the headline.

Unfortunately, because the other threads were killed (where intel was claiming magny cours was less secure because we did not have AES-NI) I don't have anything to point to.

I retract because I can't point to that thread any longer. If anyone can find an archived version I would be grateful.

[Hawkeye4077]

I remember playing Everquest and a few other games that released game updates, and for a short time the game was UNPLAYABLE on the entire line of current AMD processors at the time. At that time, I had always had an Intel computer, and I had no idea where AMD and Intel stood. But I got a laugh out of reading that at the EQ website and I still have the screenshot of the message somewhere. It's fun to see the fanboys be thrown into a position they have zero chance of defending and watch them squirm. (Typically, I've always seen MANY more AMD fanboys than Intel fanboys. Not sure if this is how it is or just luck on where I go).

If you are referring to the "runspeed/speedhacking issue" dont blame AMD, the fault was with XP. It was the OS's thread scheduler that was the problem. If you assigned the game to only 1 core it worked fine.

[flippin_waffles]

Point taken on the headline.

Unfortunately, because the other threads were killed (where intel was claiming magny cours was less secure because we did not have AES-NI) I don't have anything to point to.

I retract because I can't point to that thread any longer. If anyone can find an archived version I would be grateful.

Yeah they're pretty good at protecting intel and covering their butts around here. On the other hand I can think of a few threads where the request to be locked by AMD fell on deaf ears. Just gotta see it for what it is, and carry on. Practically everybody here knows the truth anyway, so no worries!

[duploxxx]

The source is Intel's PCNs most likely, this is not atypical as rare errata are discovered companies will issue BIOS updates in the microcode that are then routinely distributed to OEMs and ODMs. AMD does this as does Intel, this is really just someone trying to dish up dirt and blow it out of proportion.

fine, next time I won't post and you just go scratching your head further.

If I get this critical alert from HP and DELL that companies have to update bios on all series before shipment that means there is definetly something wrong....

Well seeing as who created this thread, I'm not surprised.

still pissed that i was actually able to provide data to MM on the magny , go read some more on the intel forum, that's all your able to do.

[Hornet331]

fine, next time I won't post and you just go scratching your head further.

If I get this critical alert from HP and DELL that companies have to update bios on all series before shipment that means there is definetly something wrong....

As serious as AMDs TLB errata.. that affected sooooo many users everywhere and bsods where omnipresent with this cpu..

[duploxxx]

As serious as AMDs TLB errata.. that affected sooooo many users everywhere and bsods where omnipresent with this cpu..

WTF has this got to do with old AMD errata? go screw another thread or continue to drool a bit more in the Sandy threads its IDF remember they have to bring some info... it didn't see any info that it was already able to speak

[mstp2009]

Dude, you have anger issues.

[duploxxx]

Dude, you have anger issues.

nah just sick of blue fanboys en there posting behavior so i do the same.

I posted this info without any wrong intention, info that some here on XS know that they will get a bios update for there 6 cores, just look at comments....

[Hornet331]

WTF has this got to do with old AMD errata? go screw another thread or continue to drool a bit more in the Sandy threads its IDF remember they have to bring some info... it didn't see any info that it was already able to speak

oh so now we resort to personal attacks...

Even if the big oems send out that message it hardly means anything for the enduser, as for that reason i have borught up the TLB bug for the old phenom.

If you actually would have read what i have posted without going into igonorant mode you would have seen, that this post was sarcastic, as in that the all the hubabuba about the TLB bug in the phenom was exaggerated and hardly affected any normal useres at all... (proof -> AMD section)