Real Dope on Core I7-920 DO Stepping Changes

[spylake]

I've read a lot of speculation on how a D0 differs from the C0 and C1 (yes Virginia there is a C0 and C1 stepping).

Overclocking reports seem to indicate there are thermal and or voltage improvements in the D0 stepping. On this score I have no new information other than to say Intel has not changed the electrical or thermal specs.

What Intel has announced, is a fix for a number of errata with the D0 stepping. The Core I7 family lists 102 errata. The D0 stepping fixes 16 of these and introduces one new one of it's own.

I condensed the Fixes into a word doc that you can download here

http://www.filesavr.com/corei7...rencesbetweensteppings

or for those that are leery of downloading DOC files the relevant part is shown below

Link to JPG http://xs139.xs.to/xs139/09186/errata307.jpg

 

[Idontcare]

Welcome to the forums!

The D0 stepping topic is certainly well known around here in terms of OC'ing but the specific errata fixed with the stepping has not been discussed much.

Anandtech's article on preliminary testing with D0

In your opinion, which (if any) of the fixed errata are the most critical to have been addressed?

 

[VirtualLarry]

That's a pretty curious errata that was introduced with the D0 stepping. Brand string has always been "GenuineIntel" for as long as I can remember since CPUID was introduced.

How did Intel screw that up???

 

[sgrinavi]

Interesting...

Guy comes on, first post is talking about Core I7 errata... Even has a .JPG of it all ready to DL in case we don't want to catch the AMDflu.... er a virus.

so, how do you guys spell TROLL over there?

 

[Flipped Gazelle]

Originally posted by: sgrinavi
Interesting...

Guy comes on, first post is talking about Core I7 errata... Even has a .JPG of it all ready to DL in case we don't want to catch the AMDflu.... er a virus.

so, how do you guys spell TROLL over there?

How is this a troll post? The OP mentions 16 fixes and 1 new intro'd, and also, "Overclocking reports seem to indicate there are thermal and or voltage improvements in the D0 stepping."

Sounds pretty positive from an Intel POV.

 

[Rubycon]

GIF works better for charts cut from docs.

http://pics.bbzzdd.com/users/C6FT7/35xxerr.gif

I viewed the doc (which that chart was cut from and saved as a GIF) and found a PDF link to Intel's site:

http://www.intel.com/Assets/PDF/specupdate/321333.pdf

FYI

 

[sgrinavi]

Originally posted by: Flipped Gazelle


How is this a troll post? The OP mentions 16 fixes and 1 new intro'd, and also, "Overclocking reports seem to indicate there are thermal and or voltage improvements in the D0 stepping."

Sounds pretty positive from an Intel POV.

I just find it to be a very strange first post from a guy who didn't even take the time to change his icon, has not filled out his profile and did the exact same thing on Toms Hardware BUT , hey.. whatever

 

[cusideabelincoln]

None of his posts scream, let alone remotely whisper, that he's a troll. You seem pretty defensive, though.

 

[sgrinavi]

Originally posted by: cusideabelincoln
. You seem pretty defensive, though.

LOL.. Really? Sorry -

This doesn't seem strange to anyone else?

 

[DrMrLordX]

no.

 

[Flipped Gazelle]

Originally posted by: sgrinavi

Originally posted by: cusideabelincoln
. You seem pretty defensive, though.

LOL.. Really? Sorry -

This doesn't seem strange to anyone else?

Well, we tend to regard that particular avatar as "suspicious" because spammers and trolls tend to use it. But maybe in this case the OP either was lazy or actually liked it.

But, the OP isn't criticizing anything, and hasn't even offered an opinion. Just information that pertains solely to Intel.

I'm sure most of us are so damn tired of all the AMD vs Intel nonsense that we cringe at the thought of more conflict, but so far I think we are OK here. :thumbsup:

 

[sgrinavi]

ah-ight.... sorry

 

[spylake]

Originally posted by: sgrinavi
I just find it to be a very strange first post from a guy who didn't even take the time to change his icon, has not filled out his profile and did the exact same thing on Toms Hardware BUT , hey.. whatever

My word ... Like a lot of you I was all all set to buy an I7-920 when word leaked that a D0 stepping was coming.

Some rumors said it was only a packaging change. Others said it would over clock better or would have lower power dissipation.

What was missing from the discussion was any definitive word from Intel about how the DO differed from the CO and like some of you I wanted to know before I decided which to buy.

I was in Atlanta for the FIRST Robotics Competition last month and was all set to visit MicroCenter and get a CO for a very tempting $229. I decided to wait hoping the DO rumors would be true.

So yesterday I stumble across the Intel errata document for the xeon 3520 (I7-920 by a different name) and unlike the same document for the I7-920 it lists the changes for D0 stepping. It also states that this data applies equally to the I7 family i.e. 920.

Thinking I would be appreciated for sharing some definitive answers about the DO changes I spent my precious time and crafted a word document to share with all of you at Tom's and Anandtech, which are two sites I frequent for PC info and tests.

I'm sorry I made the effort.

My friend, the fact that you twisted my innocent purpose into some sort of political AMD vs INTEL Troll boggles my mind.

You really need to chill out.

 

[spylake]

Originally posted by: VirtualLarry
That's a pretty curious errata that was introduced with the D0 stepping. Brand string has always been "GenuineIntel" for as long as I can remember since CPUID was introduced.

How did Intel screw that up???

Kinda funny I agree ... here is the extended description from Intel


AAJ69. CPUID Instruction Returns Incorrect Brand String

Problem:

When a CPUID instruction is executed with EAX = 80000002H, 80000003H and
80000004H, the return values will contain the brand string with an additional zero
between the processor number and the @ symbol. (For example, Intel Xeon CPU nnn0
@ x.xx GHz where nnn is a processor number and x.xx is the frequency)

Implication:

When this erratum occurs, the processor will report the incorrect brand string.

Workaround:

It is possible for the BIOS to contain a workaround for this erratum.

 

[daw123]

Originally posted by: spylake

Originally posted by: sgrinavi
I just find it to be a very strange first post from a guy who didn't even take the time to change his icon, has not filled out his profile and did the exact same thing on Toms Hardware BUT , hey.. whatever

My word ... Like a lot of you I was all all set to buy an I7-920 when word leaked that a D0 stepping was coming.

Some rumors said it was only a packaging change. Others said it would over clock better or would have lower power dissipation.

What was missing from the discussion was any definitive word from Intel about how the DO differed from the CO and like some of you I wanted to know before I decided which to buy.

I was in Atlanta for the FIRST Robotics Competition last month and was all set to visit MicroCenter and get a CO for a very tempting $229. I decided to wait hoping the DO rumors would be true.

So yesterday I stumble across the Intel errata document for the xeon 3520 (I7-920 by a different name) and unlike the same document for the I7-920 it lists the changes for D0 stepping. It also states that this data applies equally to the I7 family i.e. 920.

Thinking I would be appreciated for sharing some definitive answers about the DO changes I spent my precious time and crafted a word document to share with all of you at Tom's and Anandtech, which are two sites I frequent for PC info and tests.

I'm sorry I made the effort.

My friend, the fact that you twisted my innocent purpose into some sort of political AMD vs INTEL Troll boggles my mind.

You really need to chill out.

I agree with you, sgrinavi jumped to conclusions. There was nothing in the OP to suggest that it was a troll post.

Don't let this put you off from posting useful information in the future.

I found your post informative since I'm looking to buy a D0 stepping I7 CPU.

 

[spylake]

Originally posted by: daw123
[
I agree with you, sgrinavi jumped to conclusions. There was nothing in the OP to suggest that it was a troll post.

Don't let this put you off from posting useful information in the future.

I found your post informative since I'm looking to buy a D0 stepping I7 CPU.

Thanks. I'm glad someone found it useful. Good luck with your purchase, I can't find it in stock anywhere.

I've been looking at the Xeon3520 (I7-920) but Gigabyte said their boards are not certified for a Xeon while at the same time hinting they should work just fine.

If anyone knows a 3520 will work in a GA-EX58-UD4P please let me know.

 

[Idontcare]

Originally posted by: spylake

Originally posted by: sgrinavi
I just find it to be a very strange first post from a guy who didn't even take the time to change his icon, has not filled out his profile and did the exact same thing on Toms Hardware BUT , hey.. whatever

My word ... Like a lot of you I was all all set to buy an I7-920 when word leaked that a D0 stepping was coming.

Some rumors said it was only a packaging change. Others said it would over clock better or would have lower power dissipation.

What was missing from the discussion was any definitive word from Intel about how the DO differed from the CO and like some of you I wanted to know before I decided which to buy.

I was in Atlanta for the FIRST Robotics Competition last month and was all set to visit MicroCenter and get a CO for a very tempting $229. I decided to wait hoping the DO rumors would be true.

So yesterday I stumble across the Intel errata document for the xeon 3520 (I7-920 by a different name) and unlike the same document for the I7-920 it lists the changes for D0 stepping. It also states that this data applies equally to the I7 family i.e. 920.

Thinking I would be appreciated for sharing some definitive answers about the DO changes I spent my precious time and crafted a word document to share with all of you at Tom's and Anandtech, which are two sites I frequent for PC info and tests.

I'm sorry I made the effort.

My friend, the fact that you twisted my innocent purpose into some sort of political AMD vs INTEL Troll boggles my mind.

You really need to chill out.

First let's ALL chill a little folks.

Regarding sgrinavi's post, he just noticed that a certain number of typical characteristics were standing out in the OP's MO, did some homework and spotted the second identical post on THG forums (again, another typical characteristic for would-be spammer/troll) and made the suggestion here.

Why? Because he cares about the quality of these forums and he doesn't want them to degrade any more than can be helped. So let's not form a mob to get sgrinavi just because he tried doing something for the benefit of the rest of us.

spylake your information is value-add and appreciated, rest assured. As your own postcount is merely 4 posts at this time, but you openly admit enjoying these forums, that means you have been a lurker for some time now. Which means you can appreciate how information in threads is valuable to people who do not post (meaning yourself up until today).

So you should in no way judge the appreciation of your thread on the basis of merely the vocal response it receives. Check the thread's "views" vs "replies" and see how many times people have opened your thread but not mentioned anything about it.

Please continue to add value to this forum, and have some understanding of the vanguards that are in place helping keep it the kind of forum you found to be enticing enough to want to post your info in the first place.

And lastly, regarding my post at the top of the thread, in the course of your due diligence with uncovering the fixed errata is there any that stand out in your opinion as "thank god they fixed that!" type items? Since Intel waited until the D0 stepping before releasing Nehalem EP (the xeon version of i7 if you will) it makes me think there must be at least one critical errata amongst those 16 that precluded the C0/C1 steppings from being viable release candidates for Xeon markets. Your thoughts?

 

[daw123]

Originally posted by: spylake

Originally posted by: daw123
[
I agree with you, sgrinavi jumped to conclusions. There was nothing in the OP to suggest that it was a troll post.

Don't let this put you off from posting useful information in the future.

I found your post informative since I'm looking to buy a D0 stepping I7 CPU.

Thanks. I'm glad someone found it useful. Good luck with your purchase, I can't find it in stock anywhere.

I've been looking at the Xeon3520 (I7-920) but Gigabyte said their boards are not certified for a Xeon while at the same time hinting they should work just fine.

If anyone knows a 3520 will work in a GA-EX58-UD4P please let me know.

I'm after the I7 950 - for christ sake Intel, release the bloody thing.

In reply to your question, I'm not sure since the Xeon has 2 QPIs (for communication between the the two Xeon CPUs on a dual socket MB), whereas the normal I7 only has one QPI.

How come you want a Xeon chip instead of an I7? I'm not sure what benefits (if any) the Xeon has over the I7, when you are running it as a single chip.

Aigomorla is the best person to ask since he owns both. I would PM him.

 

[Rubycon]

Originally posted by: daw123

How come you want a Xeon chip instead of an I7? I'm not sure what benefits (if any) the Xeon has over the I7, when you are running it as a single chip.

Aigomorla is the best person to ask since he owns both. I would PM him.

W3520's overclock like the dickens I've been reading. Heaven know when I'll get mine!

W3520's should work in ANY X58 board with a BIOS that supports D0 stepping. Remember the 3520 is a D0!

 

[sgrinavi]

Originally posted by: spylake

You really need to chill out.

I explained myself and said I was sorry.. what more can I do?

 

[Idontcare]

Originally posted by: Rubycon

Originally posted by: daw123

How come you want a Xeon chip instead of an I7? I'm not sure what benefits (if any) the Xeon has over the I7, when you are running it as a single chip.

Aigomorla is the best person to ask since he owns both. I would PM him.

W3520's overclock like the dickens I've been reading. Heaven know when I'll get mine!

W3520's should work in ANY X58 board with a BIOS that supports D0 stepping. Remember the 3520 is a D0!

Yep all Xeon Nehalems are D0's. No C0's, no C1's. Find yourself a Nehalem Xeon and you've found yourself a guaranteed D0. Same goes for the 950's and 975's, no need to check boxes at MC and all that.

 

[dmens]

Originally posted by: VirtualLarry
That's a pretty curious errata that was introduced with the D0 stepping. Brand string has always been "GenuineIntel" for as long as I can remember since CPUID was introduced.

How did Intel screw that up???

LOL funny story actually... there is a pre-silicon test to make sure the cpuid has the correct value but for D0 nobody bothered running it. i know the guy who forgot to run that test and everyone has a laugh about the new "i7 9750".

 

[TuxDave]

Originally posted by: dmens

Originally posted by: VirtualLarry
That's a pretty curious errata that was introduced with the D0 stepping. Brand string has always been "GenuineIntel" for as long as I can remember since CPUID was introduced.

How did Intel screw that up???

LOL funny story actually... there is a pre-silicon test to make sure the cpuid has the correct value but for D0 nobody bothered running it. i know the guy who forgot to run that test and everyone has a laugh about the new "i7 9750".

Recall them all!!!!!

.... can I have them when we get them back? I need a 2nd system.

 

[daw123]

Originally posted by: Rubycon

Originally posted by: daw123

How come you want a Xeon chip instead of an I7? I'm not sure what benefits (if any) the Xeon has over the I7, when you are running it as a single chip.

Aigomorla is the best person to ask since he owns both. I would PM him.

W3520's overclock like the dickens I've been reading. Heaven know when I'll get mine!

W3520's should work in ANY X58 board with a BIOS that supports D0 stepping. Remember the 3520 is a D0!

Does the W3520 overclock any better than the I7 920?

The reason I ask is that most 920s can reach 3.8GHZ, which is a 43% overclock and not too shabby.

 

[spylake]

Originally posted by: Idontcare
And lastly, regarding my post at the top of the thread, in the course of your due diligence with uncovering the fixed errata is there any that stand out in your opinion as "thank god they fixed that!" type items? Since Intel waited until the D0 stepping before releasing Nehalem EP (the xeon version of i7 if you will) it makes me think there must be at least one critical errata amongst those 16 that precluded the C0/C1 steppings from being viable release candidates for Xeon markets. Your thoughts?

To be honest I have no clue.

I haven't seen any complaints that the I7 family is flawed in any significant area or that they crash more than they should. There has been no suggestion that with the release of the D0 stepping Engineers should stop designing bridges with the C0 variant.

I too would be curious to know how Intel chose which errata to fix and what practical effect the others have.