PDA

View Full Version : Intel Core i7 (Nehalem) also has TLB bug?


MUYA
12-01-08, 05:38 AM
Dunno if this has been discussed?

http://www.fudzilla.com/index.php?option=com_content&task=view&id=10707&Itemid=1

Remember how the Phenom sales were sorta affected by the TLB bug? Wonder if TLB bug will affect i7s sales too

Feyy
12-01-08, 05:49 AM
Ohhh, interesting! :p

Wonder if it will turn out to be as bad as the phenom one?

bacon12
12-01-08, 06:30 AM
Doubt it will effect the mighty wintel, or get as much news as the phenom bug did.

Toss3
12-01-08, 07:05 AM
"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."

So it's been fixed already :)

nekrosoft13
12-01-08, 07:43 AM
I would like to see definition of "fixed".

AMD also "fixed" it in their 9x00 series of Phenon CPUs by implementing some workaround that caused the CPU to run 10-15% slower then 9x50 series Phenon CPUs.

I have the Phenon 9550 in my server box, ok CPU, but clock per clock slower then Intel, and piss poor OC potential.

Board supports AM3, so might jump to AM3 Phenon later.

Either way, glad I haven't bought i7 yet, Board are too expensive and my quad will be good for long time ;)

hell_of_doom227
12-01-08, 08:54 AM
I would like to see definition of "fixed".

AMD also "fixed" it in their 9x00 series of Phenon CPUs by implementing some workaround that caused the CPU to run 10-15% slower then 9x50 series Phenon CPUs.

I have the Phenon 9550 in my server box, ok CPU, but clock per clock slower then Intel, and piss poor OC potential.

Board supports AM3, so might jump to AM3 Phenon later.

Either way, glad I haven't bought i7 yet, Board are too expensive and my quad will be good for long time ;)


Same here. I might pick up i920 when it comes down to $179-$199. Also, i will wait for better revision of it. Also when it comes to motherboard, it's overpriced. Waiting for that to drop down to $199-$220 at most. Afterall i might get Phenom II and AM3 motherboard...i'm sure AMD will have great pricing on it. Also, CPU just might be slightly slower then i7 or even better. The best is to wait for now. My q6700@3600 is more then enough now.

Heinz68
12-01-08, 01:28 PM
I think it's Fudzila reporting news which actually is not news to get lots of hits. The article was already linked (http://www.google.ca/search?q=Nehalem+also+has+a+TLB+bug&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a) on lots of sites and the day is not over.

Lets see if we get some official answer from Intel but I think the answer is already there in the Intel PDF document (page 37) (http://download.intel.com/design/processor/specupdt/320836.pdf) the Fud did quote.
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.

For some people who said they're glad for not buying the i7 yet, I can only say I'm glad that I didn't buy the Core 2. :) j/k

According to Wikipedia the Core 2 also had some TLB errata issue and same fixes were done. Look for the Chip Bugs on this page (http://en.wikipedia.org/wiki/Intel_Core_2).

I think Ghostbuster (http://www.xtremesystems.org/forums/showpost.php?p=3466632&postcount=46) at ExtremeSystems summarized it the best
By the way...

Here's the page of the desktop Core 2 Duo E7000 and E8000 series:
http://download.intel.com/design/processor/specupdt/318733.pdf
It has the exact same text written on page 47.

Here's the page of the desktop Core 2 Duo X6800, E6000 and E4000 series:
http://download.intel.com/design/processor/specupdt/313279.pdf
It has the exact same text written on page 71.

Here's the page of the desktop Pentium Dual Core E2000 series:
http://download.intel.com/design/processor/specupdt/316982.pdf
It has the exact same text written on page 59.

Here's the page of the Intel Celeron 200 series:
http://download.intel.com/design/processor/specupdt/31854702.pdf
It has the exact same text written on page 53.

Maybe FUD will wanna post about all of them?

My i7 965 is still in the box, waiting for the memory which should be here in couple days.
If this turns out to be real problem I have a option to return it in 30 days, but I don't think i will need to.

Redeemed
12-01-08, 01:39 PM
I do find it odd how so many looked down upon AMD and the Phenoms for having the TLB errata bug, and then complained about the "fix" they provided.

I'm running a Phenom 9600, with the "fix" enabled, and to this day there isn't a game I cannot play. The only time I'd say my Phenom appears "slow" is when encoding full length HD movies... which I hardly do. Then again, I'm new to encoding and have heard that by using h.264 it'll speed up the process significantly. I've yet to try it.

Anyhow, interesting news none-the-less. I wonder how it'll be received...

Toss3
12-01-08, 04:20 PM
I do find it odd how so many looked down upon AMD and the Phenoms for having the TLB errata bug, and then complained about the "fix" they provided.

I'm running a Phenom 9600, with the "fix" enabled, and to this day there isn't a game I cannot play. The only time I'd say my Phenom appears "slow" is when encoding full length HD movies... which I hardly do. Then again, I'm new to encoding and have heard that by using h.264 it'll speed up the process significantly. I've yet to try it.

Anyhow, interesting news none-the-less. I wonder how it'll be received...

The problem with AMD's fix was that it caused a "severe" drop in performance - Intel's reportedly doesn't.

Here's a quote from Guru3d:

"The BIOSes to workarround the TBL bug already are in place and it seems that this fix has very little effect on overall performance. AMD at the time simply switched off the entire TBL, and as to what Intel precisely did to address the issue .. is not yet known.

The BIOS update was issued on 14-11-08, before the products actually hit retail. In Intel's defense, so far everything we tested .. was 100% stable and extremely fast. But we'll report more once we hear more."

Heinz68
12-01-08, 07:55 PM
According to TechReport (http://techreport.com/discussions.x/15979) this is the response from Intel PR manager Dan Snyder

Do Core i7s have a Phenom-like TLB issue? No, says Intel
by Cyril Kowaliski — 4:11 PM on December 1, 2008

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.