Sign in with
Sign up | Sign in
Your question

AMD delays Phenom 2.4 GHz due to TLB errata

Last response: in CPUs
Share
November 18, 2007 4:23:16 PM

AMD delays Phenom 2.4 GHz due to TLB errata

http://www.theinquirer.net/gb/inquirer/news/2007/11/18/...

B2 rev K10 CPUs feature L3 cache miss flaw

By Theo Valich: Sunday, 18 November 2007, 5:29 PM

AMD WILL ONLY LAUNCH the Phenom 9500 and 9600. Even though the channel already got its hands on the Phenom 9700 (2.4 GHz) part, it will have to be pulled off from the shelves

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

In the BIG picture this is likely not a big deal.

In the short term, micropicture.. another black eye.

AMD has some serious Karma issues of late.. Talk about leaving the impression that they are the gang that can't shoot straight.
November 18, 2007 4:30:21 PM

OMFG

What a train smash.

This errata stuff is BS speak for "our 2.4 GHZ yields are crap". You dont get different errata at different speed bins.
November 18, 2007 4:56:39 PM

"This problem was found during speed-binning the B2 revision processors, and this was the cause for the Phenom FX 3.0 GHz delay. It turns out that some CPUs running at 2.4 GHz or above in some benchmarking combinations, while all four cores are running at 100% load, can cause a system freeze."

I wonder what this means for overclocking... will you get random crashes even if the CPU is tested to be 'stable'?
Related resources
November 18, 2007 5:16:10 PM

So where are the guys who said retail phenoms would easily hit 3.2 GHz OCs?
November 18, 2007 5:17:26 PM

Quote:
but it will bring serious disappointment to AMD fanboys and market itself. They are probably blubbing already.


:whistle: 
November 18, 2007 5:26:03 PM

more bad news for AMD, sigh...
more bad news for us.
November 18, 2007 5:39:47 PM

Why does the inquirer, of all places, have an inside track on this type of news? Does anyone else have difficulty in accepting the voracity of such news from such a source?
November 18, 2007 6:56:35 PM

And here I was hoping the first thing the Inq would post today was their benchmarks. Maybe NDATech as they're called will.
November 18, 2007 7:20:23 PM

onestar said:
Why does the inquirer, of all places, have an inside track on this type of news? Does anyone else have difficulty in accepting the voracity of such news from such a source?


Well the same group of people believed it when Inq reported Phenom can be easily overclocked to 3.0Ghz. :sarcastic:  :sarcastic: 

But to me, Inq is just a rumor source. Since the K8 days, Inq had always claimed to have inside source in AMD (reverse hyper-threading anyone?). Before other credible news sources report the same, I'll treat it like a rumor.
November 18, 2007 7:32:39 PM

There I was, hoping beyond hope that Phenom would be out and provide some good numbers, not expecting Phenom to be at the top, but at least close enough to worthwhile if the price was good. Sure, maybe the Inq has had its share of bad reporting, but a few other sites have not had much positive to say so far for Phenom.

Back in the hopeful department, perhaps AMD will get the problem straightened out quickly and the chips will ship before Christmas. If not, raise the glasses and toast to Intel for the holidays.
November 18, 2007 8:07:53 PM

If you read the whole paper it states that it will be fixed with a BIOS update - similar to the microcode update that Intel had for their TLB errata. Oh noes doom gloom for AMD lol.
November 18, 2007 8:13:38 PM

thematrixhazuneo said:
If you read the whole paper it states that it will be fixed with a BIOS update - similar to the microcode update that Intel had for their TLB errata. Oh noes doom gloom for AMD lol.


Actually, Inq stated that since AMD did not have built in Micro-code update like Intel, they have to pull everything from the shelf.

Quote:
Even though the channel already got its hands on the Phenom 9700 (2.4 GHz) part, it will have to be pulled off from the shelves...

...However, unlike Intel, that has a micro-code update function in all of its CPUs, AMD is forced to delay the introduction of the part.


This likely means no more Phenom 9700 for this holiday season.
November 18, 2007 8:34:16 PM

From AMD

"The Translation Lookaside Buffer (TLB) errata is an L3 protocol issue causing a system hang when running certain client workload applications independent of platform. AMD is immediately introducing an updated BIOS which will correct the TLB errata".

Either way, sucks that more bad news had to pile on....
November 18, 2007 8:53:50 PM

AMD FTW!
November 18, 2007 8:59:52 PM

thematrixhazuneo said:
From AMD

"The Translation Lookaside Buffer (TLB) errata is an L3 protocol issue causing a system hang when running certain client workload applications independent of platform. AMD is immediately introducing an updated BIOS which will correct the TLB errata".

Either way, sucks that more bad news had to pile on....


Wait until tomorrow and the stock market opens. Then, instead of an up day for AMD because of the chip line being introduced, there might be even more bad news because of this chip being recalled. I wonder if those new investors from overseas will feel like needing a drink of something strong.
November 18, 2007 9:37:04 PM

So they have to recall the chips, will we see them as tri-cores?
November 18, 2007 9:58:12 PM

xcetera said:
So they have to recall the chips, will we see them as tri-cores?


I doubt you'll see anything see, that's a characteristic of a paper launch.
November 18, 2007 10:24:54 PM

Well good thing AMD/ATI released what I consider successful mid-range video cards, or else they be dead much sooner.
November 18, 2007 10:37:10 PM

It seems the Inq isn't the only news site saying the Phenom is delayed.
http://www.dailytech.com/article.aspx?newsid=9691

Read a few paragraphs down and it states that the Phenom 2.4ghz will be launched sometime in December. I sure hope its before holidays.
November 18, 2007 10:41:03 PM

So you are willing to delay the obituary? I believe that AMD/ATI will survive, but will need some time to recover. Between the overseas investment AMD just received and the replacement of Ruiz, AMD may have some good days ahead of it. And none too soon, I might add.

Edit note: reply to Evilonigiri
November 18, 2007 10:51:57 PM

snowmanman said:
It seems the Inq isn't the only news site saying the Phenom is delayed.
http://www.dailytech.com/article.aspx?newsid=9691

Read a few paragraphs down and it states that the Phenom 2.4ghz will be launched sometime in December. I sure hope its before holidays.


Ok, if AMD gets the 2.4 ghz chip out in Dec, then there's some hope. But the pricing seems a bit high to me if these chips are supposed to take on Intel's equivalent chips. At the moment, all we can do is wait and see how things play out. At least, that's my 2 cents worth.
November 18, 2007 10:53:00 PM

Sounds to me like a heat issue:
Quote:

9500 (2.2 GHz) and 9600 (2.3 GHz) parts are unaffected by the errata. Some 9500/9600 parts may even be overclocked to 2.6, 2.8, 2.9, 3.0 GHz and they will have no problems whatsoever, while some will have this error.


So its basically saying, while 9600 and 9500 are unaffected by this errata, some chips will encounter problems when being overclocked, and some won't.

To be honest, this really deterred me from building an AMD Phenom rig. If I can't purchase a Phenom 9500, and overclock it at least to a 9700, then I don't really see why I should go with Phenom when I can get another Q6600 with the same money.

Maybe I'll have to wait for the B3 revision...which, as Baron stated earlier before, reached 4.0Ghz in lab.
November 18, 2007 11:25:41 PM

read it seriously
November 18, 2007 11:51:25 PM

This is an interesting development but maybe it explains why the L3 performance is so bad. The L3 can load directly to L1 just like main memory if I read the diagram right from real world tech.

For heavy threading you would do a lot of shared loads to different cores.
November 18, 2007 11:52:23 PM

yomamafor1 said:
Sounds to me like a heat issue:
Quote:

9500 (2.2 GHz) and 9600 (2.3 GHz) parts are unaffected by the errata. Some 9500/9600 parts may even be overclocked to 2.6, 2.8, 2.9, 3.0 GHz and they will have no problems whatsoever, while some will have this error.


So its basically saying, while 9600 and 9500 are unaffected by this errata, some chips will encounter problems when being overclocked, and some won't.

To be honest, this really deterred me from building an AMD Phenom rig. If I can't purchase a Phenom 9500, and overclock it at least to a 9700, then I don't really see why I should go with Phenom when I can get another Q6600 with the same money.

Maybe I'll have to wait for the B3 revision...which, as Baron stated earlier before, reached 4.0Ghz in lab.


Actually I believe it was 3.4GHz, but it's possible. Again the issue will be power consumption.
November 19, 2007 12:31:07 AM

AMD will pull it together...this is a minor set back....to all you trolls on this thread please grow the hell up.
November 19, 2007 12:41:55 AM

caamsa said:
AMD will pull it together...this is a minor set back....to all you trolls on this thread please grow the hell up.


No thanks. This just further illustrates AMD's inability to launch competitve products on-time. So yes, we will talk about it. And yes, we will throw AMD fanboy statements back in their faces when reality starts to prove all the wild claims and powerpoint presenations wrong.

Of course I'm sure you went over to the other two threads that try to claim Phenom is faster, but fail to mention it was put against a severly underclocked Intel processor (clock for clock) comparison. I'm sure you went over there and called them trolls, right?
November 19, 2007 12:57:06 AM

TechnologyCoordinator said:
No thanks. This just further illustrates AMD's inability to launch competitve products on-time. So yes, we will talk about it. And yes, we will throw AMD fanboy statements back in their faces when reality starts to prove all the wild claims and powerpoint presenations wrong.

Of course I'm sure you went over to the other two threads that try to claim Phenom is faster, but fail to mention it was put against a severly underclocked Intel processor (clock for clock) comparison. I'm sure you went over there and called them trolls, right?


Lets wait for the benchmarks from a reputable site. Throwing your negativity in little chunks is trolling. I am claiming nothing, I just get sick of reading a thread and having to skip all the little troll remarks. Reminds me of grade school when one kids said "my dad can beat up your dad" Let's just wait and see who beats who. This premature crap goes on way to much.

AMD has launched and will continue to launch great products. You just want to be a trouble maker...........
November 19, 2007 1:05:31 AM

Let's wait until it's actually released to make the judgements.
November 19, 2007 1:49:57 AM

I´m glad they didn't recall their phenom logo too. Now that would be something.
November 19, 2007 2:40:44 AM

BaronMatrix said:
Actually I believe it was 3.4GHz, but it's possible. Again the issue will be power consumption.


I think you actually said 4.0Ghz... :sarcastic:  :sarcastic: 

But again, I guess time will tell.
November 19, 2007 2:46:24 AM

BaronMatrix said:
This is an interesting development but maybe it explains why the L3 performance is so bad. The L3 can load directly to L1 just like main memory if I read the diagram right from real world tech.

For heavy threading you would do a lot of shared loads to different cores.


No. The L3 performance is a problem of its own, due to FIFO. David Kanters also said it in his analysis.

I think this is more of a temperature issue, and electromagnetic interference issue. If I remembered correctly, AMD puts L2 to L3 wiring directly above the die itself. This means, in order for the data to transfer from L2 to L3, it has to first go through about 10+ layers of transistors, to reach to the top of the die. During the process, it will be subjected to heat and capacitance interference. Then the wiring has to go through the top of the cores, which will be subjected to heat and electromagnetic interference. After that, the wiring has to go through another 10+ layers of transistors, and finally reached L3.

In short, diagram and real life implementations are two different things. On the die shot, Barcelona is indeed nothing short of elegant, but in real life performance, it is nothing short of a disaster.
November 19, 2007 2:47:38 AM

caamsa said:
Lets wait for the benchmarks from a reputable site. Throwing your negativity in little chunks is trolling. I am claiming nothing, I just get sick of reading a thread and having to skip all the little troll remarks. Reminds me of grade school when one kids said "my dad can beat up your dad" Let's just wait and see who beats who. This premature crap goes on way to much.

AMD has launched and will continue to launch great products. You just want to be a trouble maker...........


We'll see how AMD will handle ths situation, then we'll see if TC is indeed trolling or not.
November 19, 2007 3:00:32 AM

Now Fudzilla is saying late Q108 for 2.4Ghz.
http://www.fudzilla.com/index.php?option=com_content&ta...
Quote:

Today, on Monday the 18th AMD will finally introduce 2.2GHz Phenom 9500 and 2.3GHz clocked Phenom 9600. The 2.4GHz will be delayed until late Q1 and so will any faster Phenom. There won't be any 2.6, 2.8 or faster parts until late Q1.


Doesn't look too good if you ask me. It also seems like early stepings will not be too safe for overclocking.

Dailytech is a little more conservative, stating December for 9700
http://www.dailytech.com/AMD+Spider+Crawls+Up+the+Water...

Quote:

AMD guidance originally stated that a 2.4GHz Phenom, dubbed the Phenom 9700, would also launch on November 19, however last minute roadmap updates indicate this chip will come in December instead.


So 2.4Ghz won't launch tomorrow, that's a definitive thing. Let's see how long will AMD take to correct this TLB issue.
November 19, 2007 3:01:42 AM

a good TLB and replacement policy is vital to reduce page faults. Let them take their time.
November 19, 2007 3:36:03 AM

I am considering a new build, and don't know if I should buy now or wait.

Which do you think will be available first, Phenom or Nehalem/Gesher chips..?




/sarcasm
November 19, 2007 3:43:31 AM

caamsa said:
Lets wait for the benchmarks from a reputable site. Throwing your negativity in little chunks is trolling. I am claiming nothing, I just get sick of reading a thread and having to skip all the little troll remarks. Reminds me of grade school when one kids said "my dad can beat up your dad" Let's just wait and see who beats who. This premature crap goes on way to much.

AMD has launched and will continue to launch great products. You just want to be a trouble maker...........



Here are a few new reviews:
Quote:
Finally, we have to compare Intel’s quad-core offerings to Phenom. In all but a few of PCMark Vantage’s individual tests, the 2.4GHz Core 2 Quad Q6600 outperformed the 2.4GHz Phenom 9700. Although the Phenom did exhibit measurable IPC improvements over Athlon 64 processors, it seems the improvements aren’t enough to catch the Core 2.
Hot Hardware Review
Quote:
Put simply, AMD's best quad-core CPU last week was the Phenom 9700. Now, though, it's the Phenom 9600: AMD cannot produce effective yields at 2.4GHz. We can debate all day whether the majority of consumer software is threaded enough to take advantage of four execution cores, but the immutable fact remains that AMD's fastest quad-core offering is slower than Intel's slowest. Compounding this depressing statement for AMD is the January 2008 launch of Penryn-based Core 2 Quads, furthering Intel's performance dominance.
Hexus Review

November 19, 2007 3:55:22 AM

The Inquirer talk nonsense most of the time ;) 
November 19, 2007 4:09:40 AM

THG has it's review on the Phenom up as well.
THG review
November 19, 2007 5:16:43 AM

caamsa said:
Lets wait for the benchmarks from a reputable site.


Agreed.

Please keep in mind, I'm not some blind Intel fanboy. I'm an enthusiast that's pissed off at AMD for TOTALLY dropping the ball.

Here are my last three processors:

Opteron 175 (current system) <--- Evolution of second system
Athlon 3200+ (S939) <---Second system I built
Athlon 2800+ (Socket A, Barton core) <---First system I built

Before those were my Intel procs:
P4 1.8 Ghz
PIII 500 mhz
Pentium 100 mhz
November 19, 2007 5:49:34 AM

NMDante said:
THG has it's review on the Phenom up as well.
THG review


I think the beginning of the second sentence tells the story "After delays of more than a year". If this processor had come out a year ago, it would have been competitive to Intel's C2D and would therefore be something worth noting. As it is, its just not good enough. With the 9700 is being temporarily withdrawn, we can ignore its results. To add insult to injury, the faster AM2 chips beat it in a surprising number of games and apps. I now wonder how the QFX with an FX74 would do in comparison, as it isn't listed in the cpu charts.

Edit note: By comparing the FX74 and the FX62 from the 2006 cpu charts, it seems that the FX74 would beat the new Phenom. At least in the 3DMark06 CPU chart, the FX74 got 3792 marks, compared to Phenom 9700 at 3347 marks and the 9600 at 3245 marks. Looks like the old FX74 was better than I'd given it credit for.

Maybe with some time, AMD can get the new chips and their variations to perform. As it is, the only way I see any hope for them is if they are priced very low, so that a person might buy the platform with the hope of a better chip coming down the line in another 6 months. Even then, well, I wonder if the whole idea was worth a single thing.
November 19, 2007 5:59:56 AM

sailer said:
I think the beginning of the second sentence tells the story "After delays of more than a year". If this processor had come out a year ago, it would have been competitive to Intel's C2D and would therefore be something worth noting. As it is, its just not good enough. With the 9700 is being temporarily withdrawn, we can ignore its results. To add insult to injury, the faster AM2 chips beat in in a surprising number of games and apps. I now wonder how the QFX with an FX74 would do in comparison, as it isn't listed in the cpu charts.

Maybe with some time, AMD can get the new chips and their variations to perform. As it is, the only way I see any hope for them is if they are priced very low, so that a person might buy the platform with the hope of a better chip coming down the line in another 6 months. Even then, well, I wonder if the whole idea was worth a single thing.


I agree. It came out too late compared to the competition's CPU, and even then, it didn't beat the competition's year old 65nm based CPU. And like you mentioned, the recall of the 9700 and the fact that the 6400+ X2 beats it in certain applications, makes this even worse.

AMD can't price these too low, or they would've been better off not offering them at all. At the price point they are right now, they can be a good upgrade for existing AM2 board owners. Of course, they won't be priced at tray prices at the initial release, so it will be interesting to see how many will jump on board. For comparasion, Directron has them listed for $286 (9500) and $322 (9600). The Q6600 is listed for $278, and it beat the 9600 in almost every benchmark. Again, the price could drop, but I don't see how AMD can drop prices within a month of release, and not take a substantial hit on their margins.
a b à CPUs
November 19, 2007 8:24:47 AM

Come on AMD, RECOVER.
a b à CPUs
November 19, 2007 9:35:30 AM

Toms review was in fact the most "sympathetic" of the lot for AMD fanbois ... thats something we might take on board anyway.

I noted two which used engineering samples ... were the other reviews all using engineering samples or were some using retail?

I'm still slogging through them as they pop up.

November 19, 2007 10:07:20 AM

I have seen new benchmarks, also got one for testing lets just say they just wasted there time and money LOL
its slower that a q6600 on EVERY benchmark on every game we tried, so much for trying to compete against intel.

The benchmark speaks for them self , sure its cheap but my customers want performance, not price.

Only reason i carry amd in my store = budget systems for customers i was hoping for something with more power guess not.
!