[Intel] There is NO TLB bug in Nehalem.

Printable View

12-01-2008, 02:49 PM
qurious63ss

[Intel] There is NO TLB bug in Nehalem.

http://techreport.com/discussions.x/15979

TECH REPORT:

"While browsing news sites today, we came upon several reports saying Core i7 processors suffer from a translation lookaside buffer-related erratum. Those reports drew parallels between the TLB bug that plagued AMD's first quad-core CPUs and a note in Intel's Core i7 specification update, which states in part:

In rare instances, improper TLB invalidation may result in unpredictable system behavior, such as system hangs or incorrect data. Developers of operating systems should take this documentation into account when designing TLB invalidation algorithms. For the processors affected, Intel has provided a recommended update to system and BIOS vendors to incorporate into their BIOS to resolve this issue.

We asked Intel PR manager Dan Snyder for the chipmaker's official take on the issue, and he replied with the following:

This is simply a pointer to a previous document written in April 2007. This document is an application note (advises on programming techniques) that programmers have had since April of 2007. This item in the Nehalem spec sheet is a web pointer, under the heading "spec clarification". The reporter who wrote this did not contact us and we will try to clarify this with him.

In other words, Intel shareholders probably shouldn't be losing any sleep over this.

For reference, the AMD TLB erratum caused data corruption and system hangs in periods of high CPU utilization, and AMD halted shipments of quad-core Opterons for months because of it. On the desktop, the company released a BIOS fix that crippled the performance of early quad-core Phenoms in many apps. All Phenoms with model numbers ending in "50" and all shipping quad-core Opterons lack the erratum, though.

Update: Snyder has sent us a new statement that makes the situation even clearer:

The "AAJ1 Clarification of TRANSLATION LOOKASIDE BUFFERS" document is a SPEC CLARIFICATION, and is simply a pointer to a previous document written in April 2007.

SPEC CLARIFICATION AAJ1 was initially added due to an issue on the Intel® Core 2 Duo processor which was previously corrected with a BIOS update; this issue does not impact the Nehalem Family of CPUs. There are errata on the Intel® Core i7 processor that relate to the TLB. These all relate to improper translations or error reporting, and all of those that impact functionality have been fixed via BIOS updates prior to Core i7 launch.

(The emphasis is Intel's.)"
12-01-2008, 03:02 PM
George_o/c

No way ... If FUD claims it's true, then it is ... :ROTF: :lol: :rofl:
Can't understand the guys @ FUD ... :D
12-01-2008, 03:09 PM
qurious63ss

For clarification purposes I was going to post this in the "[FUD TLB]" thread but Fugger closed it so I had to make a new thread and if Fugger wants to close this thread too then that's fine as long as people know the truth. FUDZILLA=FAIL
12-01-2008, 03:10 PM
Zucker2k

Looks like FUD is about to eat some dirt from Intel.
12-01-2008, 03:13 PM
Shintai

FUDzilla is the reason why I applaud goverment censorship.
12-01-2008, 03:20 PM
freeloader

Quote:

Originally Posted by Shintai

FUDzilla is the reason why I applaud goverment censorship.

Too bad the same government wouldn't censor you! Joking BTW. :rofl:

The TLB is not an issue unless it causes data corruption and system crashes. No such issue here as it's been fixed. Good job by Intel catching it before releasing any CPU's to the public.
12-01-2008, 03:27 PM
Cooper

There's at least one errata for 100k transistors in every complex IC.
12-01-2008, 03:33 PM
Hornet331

Quote:

Originally Posted by Shintai

FUDzilla is the reason why I applaud goverment censorship.

Thats the reason i just skip the thread when its from fud/inq. :yepp:

As said in the other thread, i thing fuad needed some presents for christmas, that little "hoax" should have earned him a increase in hits for december. :clap:
12-01-2008, 04:01 PM
Rammsteiner

Quote:

Originally Posted by freeloader

Too bad the same government wouldn't censor you!

Fixed:p:

Anyway, it was questionable in the first place:rolleyes: Besides this clarification you could wonder whether it would have affected Intel in the same way as it did AMD, whether Intel was in the position to suffer from it or not like AMD.

Good to see this guy enlightened why this is no issue since saying 'There is no TLB bug' could have various meanings though:p:
12-01-2008, 04:01 PM
foch3 -USA-

Quote:

Originally Posted by Shintai

FUDzilla is the reason why I applaud goverment censorship.

Wow are you serious... You can't be. A lot of things start off as "FUD" then turn out to be true.

Who get's to determine whats FUD or not? Thankfully FUGGER doesn't run this website that way.
12-01-2008, 04:15 PM
qurious63ss

Quote:

Originally Posted by freeloader

Too bad the same government wouldn't censor you! Joking BTW. :rofl:

:rofl:I gotta admit that was pretty funny. :ROTF:
12-01-2008, 04:17 PM
Shintai

Quote:

Originally Posted by foch3 -USA-

Wow are you serious... You can't be. A lot of things start off as "FUD" then turn out to be true.

Who get's to determine whats FUD or not? Thankfully FUGGER doesn't run this website that way.

I think you missed point...a few times :p:
12-01-2008, 04:22 PM
ownage

Quote:

Originally Posted by foch3 -USA-

Wow are you serious... You can't be. A lot of things start off as "FUD" then turn out to be true.

Who gets to determine whats FUD or not? Thankfully FUGGER doesn't run this website that way.

Writing articles about every possible outcome creates lots of FUD and he has to be right about some things. I think its a good thing Shintai is always skeptical. Lots of guys here believe everything they want and create hundreds of BS post about new products.
A simple fact is that Fuad's articles are mostly BS (Nehalem TLB bug for example). Why is Fudzilla.com so popular while the quality is low? Because people love BS. The only thing Fuad wants is attention. Lots of page hits means more money for him. Don't you think he knew this supposed Nehalem TLB bug was BS? Off course he knew. He's laughing at all of you guys who think Fudzilla writes good articles from time to time.

Stop with the hatred towards Shintai and get a life ASAP!
Nobody in the business has respect for fudzilla, only nubs do. So yeah please go defend fudzilla lol.
12-01-2008, 04:40 PM
nn_step

Quote:

Originally Posted by Cooper

There's at least one errata for 100k transistors in every complex IC.

that point is debatable, but I do agree with the general principle given a complex enough device, there will be geometrically more errata as complexity grows.
12-01-2008, 05:06 PM
headman800

Lol Fudzilla
12-01-2008, 05:25 PM
Ghostbuster

As with the older thread.. The words "Told you so!" was pretty clear as I pointed out the same "bug" with the C2Ds... :D
12-01-2008, 05:28 PM
Zucker2k

I think the point we're missing here is how such news could trigger people selling off stocks on the market; now I'm not blaming FUD for the huge loss on WS today, but I bet there are people who dumped shares on the news. In this fragile market, such bad journalism, especially involving a critical player in the technologies futures is very irresponsible and dangerous. I mean all FUD had to do was do a little bit of investigative journalism. I guess that's too much to ask. I'm beginning to think these guys do all their journalism from behind their PCs. :shakes:
12-01-2008, 05:33 PM
flippin_waffles

Well what else would Intel say. Of course they are going to deny anything negative in the press. That's standard procedure for crying out loud. lol Deny, deny, deny. Did people expect them to raise their hand and say, "guilty as charged". Maybe there is an issue, and maybe their isn't, but i'm certainly not going buy it hook, line and sinker because of a quick PR release.
12-01-2008, 05:38 PM
Ghostbuster

Quote:

Originally Posted by flippin_waffles

Well what else would Intel say. Of course they are going to deny anything negative in the press. That's standard procedure for crying out loud. lol Deny, deny, deny. Did people expect them to raise their hand and say, "guilty as charged". Maybe there is an issue, and maybe their isn't, but i'm certainly not going buy it hook, line and sinker because of a quick PR release.

As with the other thread, I think you missed this...

Core 2 Duo E7000 and E8000 series http://download.intel.com/design/pro...pdt/318733.pdf
Same exact text written on page 47.

Core 2 Duo X6800, E6000 and E4000 series http://download.intel.com/design/pro...pdt/313279.pdf
Same exact text written on page 71.

Pentium Dual Core E2000 series http://download.intel.com/design/pro...pdt/316982.pdf
Same exact text written on page 59.

Intel Celeron 200 series http://download.intel.com/design/pro...t/31854702.pdf
Same exact text written on page 53
12-01-2008, 05:54 PM
flippin_waffles

Actually no, I never missed it at all. But like I said, i'll give it a little more time to see if anything else turns up. Something doesn't feel right as far as i'm concerned and smells a little fishy. But, we'll see.
12-01-2008, 06:00 PM
Ghostbuster

Quote:

Originally Posted by flippin_waffles

Actually no, I never missed it at all. But like I said, i'll give it a little more time to see if anything else turns up. Something doesn't feel right as far as i'm concerned and smells a little fishy. But, we'll see.

Believe what you want, as I said beforehand FUD is FUD. I've already pointed out the proof even before this verification was published.. ;)

Let me re-quote this...

Quote:

The "AAJ1 Clarification of TRANSLATION LOOKASIDE BUFFERS" document is a SPEC CLARIFICATION, and is simply a pointer to a previous document written in April 2007.

SPEC CLARIFICATION AAJ1 was initially added due to an issue on the Intel® Core 2 Duo processor which was previously corrected with a BIOS update; this issue does not impact the Nehalem Family of CPUs. There are errata on the Intel® Core i7 processor that relate to the TLB. These all relate to improper translations or error reporting, and all of those that impact functionality have been fixed via BIOS updates prior to Core i7 launch.
12-01-2008, 06:09 PM
flippin_waffles

Thanks for the permission, I believe I will! I'll wait for virtualization reviews. Maybe by that time we'll actually have some Shanghai VM articles.
12-01-2008, 06:25 PM
Spawne32

there needs to be a warning label on anything that links back to fud or inq on this site :rofl:
12-01-2008, 06:31 PM
Russian

There is also a good explanation on Theo's Blog.
http://theovalich.wordpress.com/2008...-story-at-all/
12-01-2008, 06:32 PM
qurious63ss

Quote:

Originally Posted by flippin_waffles

Actually no, I never missed it at all. But like I said, i'll give it a little more time to see if anything else turns up. Something doesn't feel right as far as i'm concerned and smells a little fishy. But, we'll see.

Sounds to me like it's wishful thinking on your part. Not a good way to cover up an issue by documenting it now is it.
12-01-2008, 07:02 PM
Scyphe

Quote:

Originally Posted by freeloader

Too bad the same government wouldn't censor you! Joking BTW. :rofl:

The TLB is not an issue unless it causes data corruption and system crashes. No such issue here as it's been fixed. Good job by Intel catching it before releasing any CPU's to the public.

Funny, that's pretty much a description of the AMD TLB errata. The difference is that they went public and took some time to sort out a fix, yet AMD bombed for it's honesty. Intel on the other hand didn't say a single word and now they're performing damage control.
12-01-2008, 07:10 PM
Warboy

I knew there wasn't really, That's why the Topic Title for it was [FUD] lol.
12-01-2008, 07:16 PM
qurious63ss

They are having to perform damage control because of FUD spread by Fudo. According to Intel, Nehalem does not have a TLB bug. C2 had a TLB errata that was disclosed as soon as it was found so not sure what you mean by them "not saying a single word".
12-01-2008, 07:16 PM
JumpingJack

Quote:

Originally Posted by Warboy

I knew there wasn't really, That's why the Topic Title for it was [FUD] lol.

:) you gotta love FUD for entertainment, for technical detail ... look elsewhere :)

You know, the spec update with the errata was published on the day Core i7 officially launched, I am actaully surprised it took as long as it did to find some back alley HW site to try to explode it into a major deal....

My standard spiel on errata -- they happen, there is no such thing as a bug free CPU. However, most any and all errata are benign, meaningless with a probability of occurence so low that many times neither AMD nor Intel have any planned fixes, most of this is due to the fact that it requires highly unusual circumstances to trigger the errata and the other is that validation testing is quite extensive.

Errata are almost always overblown by both the HW press as well as the discussions in enthusiast forums... what makes this particular revelation so appauling is that what Fuad stated wasn't even listed as a bug but a documentation clarification on the behavior of the TLB -- and it was old news, it was part of the core 2 duo spec update simply carried over.

See for example: http://download.intel.com/design/pro...pdt/313279.pdf page 71 ... exactly the same clarification (not even an errata).

Even the now famous Errata 298 TLB bug in Phenom was way way overblown, the odds of any issues for desktop usage was basically none whatsoever.

Rest assured -- any new product released and any new spec update listing errata will most certainly create a headline just for the hits, the content and severity will always be, more or less, irrelevant.

Jack
12-01-2008, 08:59 PM
kl0012

Intel's response:
The "AAJ1 Clarification of TRANSLATION LOOKASIDE BUFFERS" document is a SPEC CLARIFICATION, and is simply a pointer to a previous document written in April 2007. SPEC CLARIFICATION AAJ1 was initially added due to an issue on the Intel Core 2 Duo processor which was previously corrected with a BIOS update; this issue does not impact the Nehalem Family of CPUs. There are errata on the Intel Core i7 processor that relate to the TLB. These all relate to improper translations or error reporting, and all of those that impact functionality have been fixed via BIOS updates prior to Core i7 launch.

http://www.techreport.com/discussions.x/15979
12-01-2008, 10:33 PM
xsbb

FUD should never be taken seriously. Ever. Never ever.
12-01-2008, 11:11 PM
gOJDO

Quote:

Originally Posted by xsbb

FUD should never be taken seriously. Ever. Never ever.

WoW! What a discovery. :D
12-01-2008, 11:20 PM
Rammsteiner

Quote:

Originally Posted by xsbb

FUD should never be taken seriously. Ever. Never ever.

Hey, could you please invent a sort of square with rounded edges to put under heavy loads:p:
12-01-2008, 11:49 PM
naokaji

Quote:

Originally Posted by xsbb

FUD should never be taken seriously. Ever. Never ever.

Well, they contradict themself so often that atleast one of the 100K 2 line stories has to be true:rofl:
12-02-2008, 12:16 AM
labs23

I bet that someone must have canceled a Core i7 upgrade upon reading the news. Someone visiting fud so often, LOL.
12-02-2008, 12:50 AM
Marvin_The_Martian

Quote:

Originally Posted by Spawne32

there needs to be a warning label on anything that links back to fud or inq on this site :rofl:

Yes on both of them..

Quote:

Originally Posted by Russian

There is also a good explanation on Theo's Blog.
http://theovalich.wordpress.com/2008...-story-at-all/

Him and FUD... :rolleyes:

:rotf:

They should have their own section called 'FUD news, don't take it serieus!'
12-02-2008, 04:22 AM
Jacky

Quote:

Originally Posted by Zucker2k

I think the point we're missing here is how such news could trigger people selling off stocks on the market; now I'm not blaming FUD for the huge loss on WS today, but I bet there are people who dumped shares on the news. In this fragile market, such bad journalism, especially involving a critical player in the technologies futures is very irresponsible and dangerous. I mean all FUD had to do was do a little bit of investigative journalism. I guess that's too much to ask. I'm beginning to think these guys do all their journalism from behind their PCs. :shakes:

I really hope he can be held responsible (yeah, that is wishful thinking) to some degree. So that shareholders and/or Intel will sue his sorry bum, so that he either never returns or takes journalism a little bit more seriously aterwards
12-02-2008, 04:32 AM
Shintai

Quote:

Originally Posted by Jacky

I really hope he can be held responsible (yeah, that is wishful thinking) to some degree. So that shareholders and/or Intel will sue his sorry bum, so that he either never returns or takes journalism a little bit more seriously aterwards

In most internet journalism. And specially IT journalism. You can write anything you like and nobody can keep you accounted for it. You would need to accuse a single person for something very bad before there is even a chance.

Like if I wrote about you molesting small children and singing nazi songs while you do it. Then I can be hold to account for it. However if I write that you eat poo and are insane. Then no.

Plus the internet is more or less free anarchy.
12-02-2008, 04:47 AM
T_Flight

It's gonna happen one day. One day something is gonna get puked out over there that will attract Intel's legal Team (assuming they are not already watching the place), and all it will take is a financial hit and he'll be through. I don't even click a link to the place because it just puts money in his pocket and gives him a reason to continue. Those hit counters are not gonna buy him out of trouble when one of these big companies legal team's comes after him.
12-02-2008, 04:51 AM
Shintai

Quote:

Originally Posted by T_Flight

It's gonna happen one day. One day something is gonna get puked out over there that will attract Intel's legal Team (assuming they are not already watching the place), and all it will take is a financial hit and he'll be through. I don't even click a link to the place because it just puts money in his pocket and gives him a reason to continue. Those hit counters are not gonna buy him out of trouble when one of these big companies legal team's comes after him.

No judge, atleast in europe would do anything. So you can send as many legal teams you want. It wouldnt cost him anything.
12-02-2008, 04:52 AM
soundood

Quote:

Originally Posted by Shintai

In most internet journalism. And specially IT journalism. You can write anything you like and nobody can keep you accounted for it. You would need to accuse a single person for something very bad before there is even a chance.

this is not true at all, and is much more complex than a few lines at XS is going to justify, if anyone has written anything about a product or person that is not true, and more importantly can be proved not to be true, then that is libel.

fud is very careful to 'quote' other sources, or disect existing data, what the FUD has said is not lies or telling porkys, it has the Intel document to prove what it is saying, its how our imagination interprets this after the AMD debacle, very clever journalism if you ask me.

to OP;
well they would say that ehh? what you expect them to say,

"yes there is an erattum, and if you buy ci7, all your family will die of smallpox?"

which as all the AMD fans know is true, just hope all you intel guys have a good supply of bodybags.

i with the Fud all the way. :up:
12-02-2008, 04:55 AM
N19h7m4r3

Smallpox?

Sweet, now I need to get an i7 rig ASAP!

Collect some inheritance :D
12-02-2008, 05:06 AM
WaxCactus

thats why people shouldnt believe everything they read..
12-02-2008, 05:07 AM
soundood

Quote:

Originally Posted by WaxCactus

thats why people shouldnt believe everything they read..

exactly my point :up:

for whatits worth, i am AMD all the way, but i had a 'play' around on a mates core i7 set up, and i have to say i was very impressed, i allways found Intel rigs to lack that 'snappy' feeling, in my opinion this has gone with the ci7, very very responsive rig.

thumbs up really, pity hes not a gamer, i would love to have flung Crysis at it, with a pair of GTX280's
12-02-2008, 05:10 AM
freeloader

Quote:

Originally Posted by soundood

this is not true at all, and is much more complex than a few lines at XS is going to justify, if anyone has written anything about a product or person that is not true, and more importantly can be proved not to be true, then that is libel.

all you have to do is challenge the opinion/statement in the relevant countries courts, as demonstrated time and again by hollywood stars, Catherine Zeta Jones comes to mind, as an example.

to OP;
well they would say that ehh? what you expect them to say,

"yes there is an erattum, and if you buy ci7, all your family will die of smallpox?"

which as all the AMD fans know is true, just hope all you intel guys have a good supply of bodybags.

i with the Fud all the way. :up:

America's definition of libel is not the same in different parts of the world. Here in Canada, it's almost impossible to find someone libel and get an awarded conviction. Very rare.
12-02-2008, 05:10 AM
Ghostbuster

Quote:

Originally Posted by soundood

to OP;
well they would say that ehh? what you expect them to say,

"yes there is an erattum, and if you buy ci7, all your family will die of smallpox?"

which as all the AMD fans know is true, just hope all you intel guys have a good supply of bodybags.

i with the Fud all the way. :up:

That is un-called for. Is this how your mind works? We proved that this is a non-event and that FUD was merely posting some sensasionalist BS.. And you tell prople will die of smallpox? :down:
12-02-2008, 05:23 AM
Hornet331

Quote:

Originally Posted by Ghostbuster

That is un-called for. Is this how your mind works? We proved that this is a non-event and that FUD was merely posting some sensasionalist BS.. And you tell prople will die of smallpox? :down:

but only people who buy Ci7. :rofl:
12-02-2008, 05:29 AM
Oliverda

There are TLB bugs in Nehalem:

AAJ42 X Plan Fix: Incorrect TLB Translation May Occur After Exit From C6

AAJ59 X Plan Fix: Writes to IA32_CR_PAT or IA32_EFER MSR May Cause an Incorrect ITLB Translation

AAJ54 X No Fix: Core C6 May Clear Previously Logged TLB Errors

source: 10-12 pages
12-02-2008, 05:31 AM
soundood

Quote:

Originally Posted by freeloader

America's definition of libel is not the same in different parts of the world. Here in Canada, it's almost impossible to find someone libel and get an awarded conviction. Very rare.

as i said, this is far more complex than a few lines at XS is going to justify, individual countries have limited laws to no laws concerning this subject,

but it seems there are 'some' cases in canada that go through ok,

TORONTO, Jan. 18 /CNW/ - A libel case brought by the Honourable Justice Dianne Pettit Baig, stemming from a February 8, 2000 article in the Globe and Mail and other media, has been settled. The defendants, who include The Globe and Mail, The Canadian Press, The Fort Frances Times, seven Fort Frances lawyers (Clare Brunetta, Ian McLennan, Wesley Derksen, Ken Koprowski, Lawrence Phillips, Emery Ruff and Donald Taylor), two members of the Fort Frances court staff (Margaret Katona and Donna Anderson) and a Fort Frances locksmith (Ken Rogoza) have all contributed to the settlement totalling $568,500.00. The amount is paid to Justice Baig for damages, including legal costs. The Globe and Mail paid $200,000.00 of the settlement amount. All of the defendants have made written apologies to Justice Baig who is a judge of the Ontario Court of Justice in the Thunder Bay and Rainy River Districts. For further information: contact her lawyer: Robert Rueter at Rueter Scargall Bennett LLP, (416) 869-3363. .

Quote:

Originally Posted by soundood

to OP;
well they would say that ehh? what you expect them to say,

"yes there is an erattum, and if you buy ci7, all your family will die of smallpox?"

which as all the AMD fans know is true, just hope all you intel guys have a good supply of bodybags.

i with the Fud all the way. :up:

Quote:

Originally Posted by Ghostbuster

That is un-called for. Is this how your mind works? We proved that this is a non-event and that FUD was merely posting some sensasionalist BS.. And you tell prople will die of smallpox? :down:

it seems the irony of my post was lost in your non humour day, lighten up dude :up:
12-02-2008, 05:37 AM
[XC] riptide

Quote:

Originally Posted by LOE

commie

Would you like more infraction points already? :rolleyes:

EDIT: You guys are unbelievable... can you not just have a discussion without the sniping and BS. Seriously? Everyday I am amazed at the caveman attitude of people in these threads. I MIGHT understand if the content of this thread was something really important! Like [enter market leader of any harware type] going out of business or something... but this?

Remember we're again disucussing a 'misleading interpretation' from Fuad. Not the end of the world.
12-02-2008, 05:40 AM
kl0012

Quote:

Originally Posted by Oliverda

There are TLB bugs in Nehalem:

AAJ42 X Plan Fix: Incorrect TLB Translation May Occur After Exit From C6

AAJ59 X Plan Fix: Writes to IA32_CR_PAT or IA32_EFER MSR May Cause an Incorrect ITLB Translation

AAJ54 X No Fix: Core C6 May Clear Previously Logged TLB Errors

source: 10-12 pages

That was already covered by Intel's response. See first post.
12-02-2008, 05:47 AM
Ghostbuster

Quote:

Originally Posted by soundood

it seems the irony of my post was lost in your non humour day, lighten up dude :up:

Urmm... Sorry I do not call that humour in any sense... especially the line "all your family will die of smallpox", sounds more like cursing. :nono:

EDIT: Telling people's family to go and die a slow and horrible death is NOT funny and its certainly NOT humour! You can compare this kind of statements with those of terrorists and extremist threats. Seriously! :mad:
12-02-2008, 05:47 AM
Oliverda

Quote:

Originally Posted by kl0012

That was already covered by Intel's response. See first post.

Intel just talked about the AAJ1 erratum. :rolleyes: These erratums are still exist. See the official Intel PDF.
12-02-2008, 05:53 AM
Hornet331

Quote:

Originally Posted by Oliverda

Intel just talked about the AAJ1 erratum. :rolleyes: These erratums are still exist. See the official Intel PDF.

Guess how many TLB "issus" are still on 65nm C2 or X2. :rolleyes:
12-02-2008, 05:53 AM
kl0012

Quote:

Originally Posted by Oliverda

Intel just talked about the AAJ1 erratum. :rolleyes: These erratums are still exist. See the official Intel PDF.

No. They talked about clarification AAJ1. And then other erratas were mentioned.
SPEC CLARIFICATION AAJ1 was initially added due to an issue on the Intel® Core 2 Duo processor which was previously corrected with a BIOS update; this issue does not impact the Nehalem Family of CPUs. There are errata on the Intel® Core i7 processor that relate to the TLB. These all relate to improper translations or error reporting, and all of those that impact functionality have been fixed via BIOS updates prior to Core i7 launch.
12-02-2008, 06:06 AM
Oliverda

Quote:

Originally Posted by Hornet331

Guess how many TLB "issus" are still on 65nm C2 or X2. :rolleyes:

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D
12-02-2008, 06:09 AM
Hornet331

Quote:

Originally Posted by Oliverda

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D

reading comprehension dude... i said TLB isuss, not erratas, also Core 2 Duo and Athlon X2. :rofl:
12-02-2008, 06:10 AM
Rammsteiner

Quote:

Originally Posted by Oliverda

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D

Whoa, epic. Nice find:p:
12-02-2008, 06:12 AM
qurious63ss

Quote:

Originally Posted by Oliverda

There are TLB bugs in Nehalem:

AAJ42 X Plan Fix: Incorrect TLB Translation May Occur After Exit From C6

AAJ59 X Plan Fix: Writes to IA32_CR_PAT or IA32_EFER MSR May Cause an Incorrect ITLB Translation

AAJ54 X No Fix: Core C6 May Clear Previously Logged TLB Errors

source: 10-12 pages

Good find!!! If I was you I would send every errata on Ci7 I find to FUDZILLA so that he can make an article on it:rolleyes: We have to find something wrong with that pesky Ci7 before it makes it to servers.;)
12-02-2008, 06:14 AM
Shintai

Quote:

Originally Posted by Oliverda

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D

Quote:

There may be missing errata numbers. Errata that have been resolved from early revisions of
the processor have been deleted, and errata that have been reconsidered may have been
deleted or renumbered.

Its not entirely the same. Plus nor Intel or AMD actually list all erratas. They list what is "needed".
12-02-2008, 06:14 AM
kl0012

Quote:

Originally Posted by Oliverda

It has 42 erratums. The Shanghai/Deneb has less.

How do you know that? Specification update for Shanghai wasn't released yet.
12-02-2008, 06:17 AM
Particle

I don't get it. The Intel guy says "there is no TLB bug" and then explicitly describes an issue where corruption occurs and states that it was fixed with a BIOS update prior to launch.

...so you're saying there's a TLB bug, but it's alright since you patched it with software? That sounds familiar.
12-02-2008, 06:17 AM
Hornet331

Quote:

Originally Posted by kl0012

How do you know that? Specification update for Shanghai wasn't released yet.

guess we have a new magic crystal ball user. :ROTF:
12-02-2008, 06:18 AM
qurious63ss

Quote:

Originally Posted by kl0012

How do you know that? Specification update for Shanghai wasn't released yet.

Well simple AMD=good, Intel=bad. Silly wabbit:D
12-02-2008, 06:18 AM
Shintai

Quote:

Originally Posted by Oliverda

There are TLB bugs in Nehalem:

AAJ42 X Plan Fix: Incorrect TLB Translation May Occur After Exit From C6

AAJ59 X Plan Fix: Writes to IA32_CR_PAT or IA32_EFER MSR May Cause an Incorrect ITLB Translation

AAJ54 X No Fix: Core C6 May Clear Previously Logged TLB Errors

source: 10-12 pages

Just as a note: C6 state means the system is in very deep sleep. PLLs, caches etc is turned of. And system state is in main memory only.

I would claim this is a non issue until i7 laptops are released.

http://images.anandtech.com/reviews/...ve/cstates.jpg
12-02-2008, 06:20 AM
Donnie27

Quote:

Originally Posted by naokaji

Well, they contradict themself so often that atleast one of the 100K 2 line stories has to be true:rofl:

Best line in the thread! Unfortunately I disagree. IMHO, I think FUD knows exactly what he's doing. I call it "Hit Phishing" or "Hit Mining". He knows sensationalism and outrageous-ness gets hits. Hits brings in the Bucks. Sure, there's enough folks hoping, wishing and praying Intel slips up to give more hits. Then there are the folks going there to argue the other side and so on. But shame on anyone calling themselves a n Xtreme-Systems Geek or Enthusiast even clicking on that site (or the other) more than once a week and then only for laughs. Folks flocking there for either Pro or Con only encourages him to continue to post such crap-tacular BS!
12-02-2008, 06:24 AM
Oliverda

Quote:

Originally Posted by Shintai

Just as a note: C6 state means the system is sleeping. PLLs, caches etc is turned of. And system state is in main memory only.

http://images.anandtech.com/reviews/...ve/cstates.jpg

I know, but for example the AAJ59 is not a C6 state related erratum.

Quote:

Originally Posted by kl0012

How do you know that? Specification update for Shanghai wasn't released yet.

You'll see it. :)
12-02-2008, 06:27 AM
Shintai

Quote:

Originally Posted by Particle

I don't get it. The Intel guy says "there is no TLB bug" and then explicitly describes an issue where corruption occurs and states that it was fixed with a BIOS update prior to launch.

...so you're saying there's a TLB bug, but it's alright since you patched it with software? That sounds familiar.

The issue is people use overall macro generalizations for terms they dont understand. So anything with bug and TLB must mean the same as AMDs TLB bug. Tho they may not have a single bit in common besides regarding to the TLB.
12-02-2008, 06:34 AM
JumpingJack

Quote:

Originally Posted by Oliverda

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D

This will slowly degrade into another flame war... AMD has not published errata for Shanhai/Deneb, Barcleona had at one point over 300 errata ... the number of errata does not make or break a product. AMD removes errata entry if there is no issue or the internalize it ... the index them sequentially, the last errata entry is number 355 ... hence, a total of 355 errata were discovered since the first stepping.

These type of problems are discovered after validation, which is pretty extensive from the commentary Torvals suggests. This is just a bunch of hay over nothing.
12-02-2008, 06:35 AM
Donnie27

Quote:

Originally Posted by qurious63ss

Well simple AMD=good, Intel=bad. Silly wabbit:D

QFT!:yepp::rofl:
12-02-2008, 06:35 AM
JumpingJack

Quote:

Originally Posted by Particle

I don't get it. The Intel guy says "there is no TLB bug" and then explicitly describes an issue where corruption occurs and states that it was fixed with a BIOS update prior to launch.

...so you're saying there's a TLB bug, but it's alright since you patched it with software? That sounds familiar.

The point in question was a spec clarification, this is not a bug ... the errata list other bugs that are either benign or have been addressed with microcode updates. ... not terribly difficult to understand.
12-02-2008, 06:39 AM
qurious63ss

Quote:

Originally Posted by Oliverda

You can count: http://www.amd.com/us-en/assets/cont...docs/41322.pdf :)

It has 42 erratums. The Shanghai/Deneb has less.

So, the i7 has exactly 77 erratums. Ohh man...maybe that is the source of the brand i7. :D

Barcelona could of just had 1 errata, TLB bug, and C2 could of had 100s but still barcelona's was a show stopper and that's the difference. Point being that you can't just count errata and say this process/arch is better then the other one it just doesn't work that way.
12-02-2008, 06:41 AM
freeloader

Quote:

Originally Posted by Shintai

The issue is people use overall macro generalizations for terms they dont understand. So anything with bug and TLB must mean the same as AMDs TLB bug. Tho they may not have a single bit in common besides regarding to the TLB.

Well it's apparent they shared a major thing in common. Data corruption and system hangs.
12-02-2008, 06:47 AM
qurious63ss

Quote:

Originally Posted by freeloader

Well it's apparent they shared a major thing in common. Data corruption and system hangs.

Yes seems that the barcelona and c2 TLB bug could of caused the same issues. Why it wasn't a big deal on C2 I really don't know maybe no one uses virtual on c2 but again they have not found a TLB bug on nehalem. As jack already mentions this is a spec clarification for something that was found on C2 and already fixed.
12-02-2008, 06:57 AM
Jacky

Quote:

Originally Posted by LOE

You mean you understand? Technically? Care to explain?

Quote:

Originally Posted by JumpingJack

This will slowly degrade into another flame war...

Wow, you have predicted the future, it's already happening. :D Please tell me should I buy AMD or Intel stock or Mcdonald's? :yepp:

qurious63ss, I don't think you understand the difference between "showstopper bug" and errata... if Intel's CPU are buggy people *will* notice and they *will* have to recall or replace their CPUs the way AMD did.
12-02-2008, 07:19 AM
Particle

Quote:

Originally Posted by JumpingJack

The point in question was a spec clarification, this is not a bug ... the errata list other bugs that are either benign or have been addressed with microcode updates. ... not terribly difficult to understand.

The odd part is that you don't seem to understand that I -do- understand yet still feel this way.
12-02-2008, 07:20 AM
qurious63ss

Quote:

Originally Posted by Jacky

Wow, you have predicted the future, it's already happening. :D Please tell me should I buy AMD or Intel stock or Mcdonald's? :yepp:

qurious63ss, I don't think you understand the difference between "showstopper bug" and errata... if Intel's CPU are buggy people *will* notice and they *will* have to recall or replace their CPUs the way AMD did.

A showtopper bug is an errata that's sever enough to make a recall because there is now workaround. Anyways, my point was that it even if a chip has 100s of errata it only takes 1 errata for the "show to stop" So you can't just compare chips by the number of errata but the severity of the errata.
12-02-2008, 07:23 AM
Shintai

Quote:

Originally Posted by freeloader

Well it's apparent they shared a major thing in common. Data corruption and system hangs.

Just to try and keep it simple. I will just paste the short descriptions.

Quote:

AMD TLB bug:
Under a highly specific and detailed set of conditions, an internal resource livelock may occur between a TLB reload and other cached operations.

Quote:

Intel TLB bug 1:
Under certain conditions when C6 and two logical processors on the same core are enabled on a processor, an instruction fetch occurring after a logical processor exits from C6 may incorrectly use the translation lookaside buffer (TLB) address mapping belonging to the other logical processor in the processor core.

Quote:

Intel TLB bug 2:
Following an exit from core C6, previously logged TLB (Translation Lookaside Buffer) errors in IA32_MCi_STATUS may be cleared.
Implication: Due to this erratum, TLB errors logged in the associated machine check bank prior to core C6 entry may be cleared. Provided machine check exceptions are enabled, the machine check exception handler can log any uncorrectable TLB errors prior to core C6 entry. The TLB marks all detected errors as uncorrectable.
Workaround: As long as machine check exceptions are enabled, the machine check exception handler can log the TLB error prior to core C6 entry. This will ensure the error is logged before it is cleared.

Quote:

Intel TLB bug 3:
Under certain conditions, writes to IA32_CR_PAT (277H) or IA32_EFER (C0000080H) MSRs may result in an incorrect ITLB (instruction translation lookaside buffer) translation.

None of the 2 companies actually writes data corruption btw.

First of all you can see that the descriptions are very very different from one another. That itself should show its not something that is related.

Bug 1 for Intel also requires the CPU to be in deep sleep, and have HT on. Both bug 1 and 3 relates to ITLB actually.
Bug 2 you would never see when you run Windows etc. Since machine checks are enabled anyway.

So as with all the other bugs on the AMD and Intel CPUs. Only 1 bug was critical. And that was the AMD TLB bug sofar. All the other 300 phenom bugs and sofar 77 i7 bugs and some 200+ C2 bugs are non critical. But that doesnt make people click on your website...does it? ;)

Just as a fun note from the Phenom list:

Quote:

Memory Instability After PWROK Assertion

For the senstionalist/clueless person. It sounds like something major. Memory instability!!!

Or this one:

Quote:

A DIV Instruction Followed Closely By Other Divide Instructions May Yield Incorrect Results

OMG..it cant calculate correctly!

Quote:

Processor Core May Execute Incorrect Instructions on Concurrent L2 and Northbridge Response

It doesnt even know what to execute and when!

Easy to make something big out of nothing if you want to. And/or doesnt understand what you are reading.
12-02-2008, 07:34 AM
Drwho?

Let's put an end to this. There is no commercial Core i7 having any issue what so ever with TLBs, the performance showed by all the reviewer is correct. The Errata quoted were about Core 2, in April 2007 ...

That is an old stuff that was putted in the Documentation, nothing to worry about.
12-02-2008, 07:41 AM
qurious63ss

Hey Francois a bit off topic, but any idea when those Core i7s are heading to the chip loaner program?
12-02-2008, 07:45 AM
Drwho?

Quote:

Originally Posted by qurious63ss

Hey Francois a bit off topic, but any idea when those Core i7s are heading to the chip loaner program?

I have no clue ... you can get a Gateway machine for 990$ :) with a 920 ...
12-02-2008, 07:55 AM
qurious63ss

Sorry I thought you knew what chip loaner program is. It's a program we have in Fab32 11X and I think Oregon has it too where we get a chip for a year to use. Currently I have Q9550 through this program but in January I will be elgible for a new chip. Would love to get a Ci7.
12-02-2008, 08:19 AM
Ziwro

Quote:

Originally Posted by Drwho?

Let's put an end to this. There is no commercial Core i7 having any issue what so ever with TLBs, the performance showed by all the reviewer is correct. The Errata quoted were about Core 2, in April 2007 ...

That is an old stuff that was putted in the Documentation, nothing to worry about.

Intel :slapass: Fuad
12-02-2008, 10:21 AM
Rammsteiner

Quote:

Originally Posted by Ziwro

Intel :slapass: Fuad

Thus far it seems to be vice versa, so far FUD managed to do anything anyway:rolleyes:
12-02-2008, 10:33 AM
Shadowmage

Quote:

Originally Posted by JumpingJack

This will slowly degrade into another flame war... AMD has not published errata for Shanhai/Deneb, Barcleona had at one point over 300 errata ... the number of errata does not make or break a product. AMD removes errata entry if there is no issue or the internalize it ... the index them sequentially, the last errata entry is number 355 ... hence, a total of 355 errata were discovered since the first stepping.

These type of problems are discovered after validation, which is pretty extensive from the commentary Torvals suggests. This is just a bunch of hay over nothing.

That's true. Wait till you guys see how many errata GT200 has :up:

I think that the worst was NV40, which made the software driver team completely hate the hardware team :lol:
12-02-2008, 11:04 AM
freeloader

Quote:

Originally Posted by Shintai

Just to try and keep it simple. I will just paste the short descriptions.

None of the 2 companies actually writes data corruption btw.

First of all you can see that the descriptions are very very different from one another. That itself should show its not something that is related.

Bug 1 for Intel also requires the CPU to be in deep sleep, and have HT on. Both bug 1 and 3 relates to ITLB actually.
Bug 2 you would never see when you run Windows etc. Since machine checks are enabled anyway.

So as with all the other bugs on the AMD and Intel CPUs. Only 1 bug was critical. And that was the AMD TLB bug sofar. All the other 300 phenom bugs and sofar 77 i7 bugs and some 200+ C2 bugs are non critical. But that doesnt make people click on your website...does it? ;)

Just as a fun note from the Phenom list:

For the senstionalist/clueless person. It sounds like something major. Memory instability!!!

Or this one:

OMG..it cant calculate correctly!

It doesnt even know what to execute and when!

Easy to make something big out of nothing if you want to. And/or doesnt understand what you are reading.

Personally I don't give a :banana::banana::banana::banana: about erratas or CPU problems from any of the companies. I just want the CPU to work correctly. If that means you have to patch via microcode update or BIOS, so be it.

BTW...if your system hangs, you'll have a good chance of data corruption. But you know that already. They go hand in hand together. You can argue semantics all you like, but the point is Intel had a bug that could've potentially caused system hangs and data corruption. They fixed it. That's the bottom line.

All times are GMT -8. The time now is 06:46 AM.

XtremeSystems