Phenom issues.. Has it been solved?

B

blackwinterday

Limp Gawd
Joined
Mar 2, 2007
Messages
145
From what i've understood there's a bug on Phenom cpu's and you would need a bios update to fix it. However it costs around %10 peformance drop. If this is the case and there's no other workaround this is a scandalous mistake by AMD and they should accept RMAing all the chips. I feel too bad about them.:(
 
It's blown way out of proportion. Very few end users would run into it. You almost have to be looking for it to find it. For me (and I do have a Phenom..), the problem is virtually non existent.
 
So what's the fuss over the net about it? Is it some kind of dirty plan to kill AMD? I thought it was a major problem on all Phenom cpu's. Dammit!:mad:
 
Idk what the fuss is about.. word on the internets is that AMD might be pushing back the launch of the B3 Phenoms to Q2.. :(
 
blackwinterday said:
From what i've understood there's a bug on Phenom cpu's and you would need a bios update to fix it. However it costs around %10 peformance drop. If this is the case and there's no other workaround this is a scandalous mistake by AMD and they should accept RMAing all the chips. I feel too bad about them.:(
Click to expand...

blackwinterday said:
So what's the fuss over the net about it? Is it some kind of dirty plan to kill AMD? I thought it was a major problem on all Phenom cpu's. Dammit!:mad:
Click to expand...
Read Kyle's take on the whole TLB issue here: http://www.hardforum.com/showpost.php?p=1031824555&postcount=54
 
blackwinterday said:
So what's the fuss over the net about it? Is it some kind of dirty plan to kill AMD? I thought it was a major problem on all Phenom cpu's. Dammit!:mad:
Click to expand...

I agree with mpcamer1220; it's blown ridiculously out of proportion! I've had my Phenom 9600 for two weeks and have yet to experience a problem that wasn't related to the in-house idiot (me!). And even then the problem was a missing driver I forgot. Cleared that up and the 790FX + Phenom 9600 combo are tons of fun and running great!

I think the fuss is all about simple fanboyism... It's so easy to take the well-worn path and then declare yourself a DIY Magellan. But now that the launch dust has settled, it turns out that Phenom is a very strong processor after all and (gasp...) affordable at that. That threatens some people, so the knee-jerk reaction is to make a fuss. Quad-core for the masses? There goes the quad-core neighborhood! :D
 
mpcamer1220 said:
Idk what the fuss is about.. word on the internets is that AMD might be pushing back the launch of the B3 Phenoms to Q2.. :(
Click to expand...
Are you serious? Where did you hear this and what was the supposed reason for the delay?
 
Pepsiennis said:
I agree with mpcamer1220; it's blown ridiculously out of proportion! I've had my Phenom 9600 for two weeks and have yet to experience a problem that wasn't related to the in-house idiot (me!). And even then the problem was a missing driver I forgot. Cleared that up and the 790FX + Phenom 9600 combo are tons of fun and running great!

I think the fuss is all about simple fanboyism... It's so easy to take the well-worn path and then declare yourself a DIY Magellan. But now that the launch dust has settled, it turns out that Phenom is a very strong processor after all and (gasp...) affordable at that. That threatens some people, so the knee-jerk reaction is to make a fuss. Quad-core for the masses? There goes the quad-core neighborhood! :D
Click to expand...
Fanboys are responsible for blowing this out of proportion? Try AMD themselves:

TechReport said:
In order to better understand this problem, TR spoke with Michael Saucier, Desktop Product Marketing Manager at AMD. Saucier confirmed that the TLB erratum can cause the system to hang when the chip is experiencing high utilization. AMD has stated previously that virtualization workloads can lead to this problem, but Saucier clarified that other workloads can trigger system hangs, as well. He characterized the issue as a race condition in the TLB logic "where the other guy wins who isn't supposed to win," and said the likelihood of the erratum causing a system hang is extremely rare.
Click to expand...
Source: http://techreport.com/discussions.x/13724

So on one hand he's saying that it will hang whenever it is under heavy load, and on the other he's saying the chances of encountering it are slim. Combine that with a rumor saying that AMD found the issue severe enough to stop shipping Barcelona, and, yeah, people are going to be wary of the chip.
 
mpcamer1220 said:
It's blown way out of proportion. Very few end users would run into it. You almost have to be looking for it to find it. For me (and I do have a Phenom..), the problem is virtually non existent.
Click to expand...
Awesome! Call all the major OEMs and let them know they can start shipping servers and desktops to customers. All this fuss for nothing, it's just been "blown way out of proportion." :p

Somehow the proportion seems justified, considering the effect it has had on shipping desktops and servers for AMD's biggest customers.
 
Have you noticed how whenever there is a swipe to take at the Phenom, everyone quotes the same source: Tech Report? I don't suppose it has upped their page hits any. Look, AMD's Saucier said, "Can... blah, blah, blah... can... blah, blah, blah... and can... blah, blah, blah." And this was well after Tech Report got their torches lit, nooses tied, and the villagers marching to Castle Sunnyvale. He never said 'did', so it's best not to take people out of context. Speaking of context, as an issue, it is much more relevant to servers than desktop. Look, it might not be called fanboyism, but Techreport's vehemence on this story is beyond passionate journalism and fast bearing down upon 'agenda'.
 
Then either AMD themselves blew this out of proportion, or TR is in the business of making up AMD quotes out of thin air.

I also don't see how people running around with their fingers in their ears about the whole thing are any better than the people blowing it out of proportion. There is an issue. What can be debated at this point is its severity. AMD is contradicting themselves left and right on this front. On one hand they stop shipments to OEMs, and on the other they provide a way to disable the fix and say you can do so with impunity. Which message am I supposed to believe?

In the face of such uncertainty, I'm leaning toward "wait for B3" as my answer, since its arrival will render the issue moot.
 
LstOfTheBrunnenG said:
... There is an issue. What can be debated at this point is its severity. AMD is contradicting themselves left and right on this front. On one hand they stop shipments to OEMs (Of server chips), and on the other they provide a way to disable the fix and say you can do so with impunity (for desktop chips). Which message am I supposed to believe?

In the face of such uncertainty, I'm leaning toward "wait for B3" as my answer, since its arrival will render the issue moot.
Click to expand...

As for severity, I put my money where my logic is and got a Phenom 9600. Nothing has locked up, fallen over, caught fire, or started calling me "Dave". So far it's a great chip that at stock performs so similarly to a Q6600 you really couldn't guess otherwise.
 
Pepsiennis said:
As for severity, I put my money where my logic is and got a Phenom 9600. Nothing has locked up, fallen over, caught fire, or started calling me "Dave". So far it's a great chip that at stock performs so similarly to a Q6600 you really couldn't guess otherwise.
Click to expand...
I run virtualization software, HTTP servers, and FTP servers on my desktop routinely for various reasons. I don't see why I, or anyone else who may run such applications in the future should have to take the chance.

Phenom runs great for you, and I'm genuinely happy that you haven't encountered any problems. I would have been one of the first to stand up cheering if Phenom had launched bug-free, and run cooler and faster than Core 2. Unfortunately that's not what's happened. While it works fine for you, I don't see why I or anyone else should be in a rush to upgrade to Phenom, especially when a new revision that will fix some (hopefully most) of these issues is forthcoming.
 
blackwinterday said:
Phenom issues.. Has it been solved?QUOTE]

TOTALLY!

http://www.theinquirer.net/gb/inquirer/news/2007/12/24/report-claims-amd-phenoms

AMD refuses to comment

By INQUIRER staff: Monday, 24 December 2007, 10:37 AM

TAIWANESE WIRE Digitimes reports AMD partners claiming high binned versions of its Phenom quad cores will now be pushed back into the second quarter of 2008.

There’s no confirmation of this from AMD, but Digitimes speculates that the real question is the future of the firm’s much vaunted “triple cores”.

The wire claims that AMD is still having difficulties sorting the translation lookaside buffer bug.
Click to expand...
 
Just use your common sense...AMD has to tell people, especially corporate customers about any known bugs or issues as credibility is extremely important in that sector. Just imagine if they had covered up that error, even if the problem was rare. Nobody would ever trust them again. Intel has the better product right now, but AMD did the right thing by bringing this error to light and not downplaying it.

So yeah, if I were a corporate buyer I would not touch Barcelona/Phenom until later this year (2Q 08 by the looks of it now)...for home use (ie. games, light video editing, pr0n, etc..) the Phenom on the new platform would be just fine.
 
supergrover said:
Just use your common sense...AMD has to tell people, especially corporate customers about any known bugs or issues as credibility is extremely important in that sector. Just imagine if they had covered up that error, even if the problem was rare. Nobody would ever trust them again. Intel has the better product right now, but AMD did the right thing by bringing this error to light and not downplaying it.

So yeah, if I were a corporate buyer I would not touch Barcelona/Phenom until later this year (2Q 08 by the looks of it now)...for home use (ie. games, light video editing, pr0n, etc..) the Phenom on the new platform would be just fine.
Click to expand...

Exactly! Well said!
 
supergrover said:
but AMD did the right thing by bringing this error to light and not downplaying it.
Click to expand...
LOL You must have been in an alternate reality where AMD did not obscure the issue and silently stop shipping CPUs. Did AMD document that erratum where you were? They gave lame excuses on this planet. :p

The severity of the bug and reaction to it can be challenged, but AMD did not handle it well at all. Allegedly the bug was caught by a customer, an unnamed large OEM.
 
Pepsiennis said:
Look, it might not be called fanboyism, but Techreport's vehemence on this story is beyond passionate journalism and fast bearing down upon 'agenda'.
Click to expand...


Utter and complete BS.Its nothing of the kind.AMD as far as I have heard,were not the ones to make the press and thier partners aware of the TRLB issue at first.It was a third party.

Edit: pxc beat me to it.

As to availability I am not the CEO of Newegg,but I do know a few people who work for Ingram Micro (I have known both since the late 80's).Both of them have told me,repeatedly,that Barcelona and Phenom,have been scarce at best.
This is why I do beleive there was a 'Stop Ship' order placed by AMD,among other things.

For anyone thats interested, NCIX is selling the 9500 for a real steal right now,below 190 ! Cdn

If I had not already ordered a 9600 BE I would have jumped on it.
 
Hey Manny! Well, it's not BS. Tech Report went overboard with their reporting. So much so that Scott Watson even issued an apology in intimating that AMD had actively deceived press and public. In the same apology he mentions that AMD limited (not stopped, as so many here hyperbolically state) shipments of Opterons to orders already received. I know that here in China, the Phenom is a very hot product, which is interesting to note as the Chinese computer DIYer is terribly careful with his purchases. If the chip had a problem it wouldn't sell here, period. But it's a hot (and available) commodity here in China. I'm sure the K10's erratum will bunghole somebody out there, but for the other 99%+ of us, the erratum will never darken our desktop, so I think some sane perspective is in order. perspective lost, as Scott Watson himself admitted.
 
The issue doesn't really affect many people and the 'fixed' versions aren't expected till early 08.
 
The problem with this bug is its extremely random behaviour.
People can come here and shout that they have a Phenom, and never have a problem... but that's not the point.
They may have a particular CPU that doesn't suffer from the bug (at stock speeds).
It's closely related to overclocking... Not all CPUs overclock equally far, at some point a critical component will fail and the CPU becomes unstable. So when someone says "My CPU overclocks to 4 GHz perfectly stable", do we assume that all CPUs of that same model are guaranteed to overclock to 4 GHz completely stable? No we don't, it's all a game of chance.

So yes, I'm sure there are CPUs that will rarely, or even never exhibit this particular bug. But I'm also sure that some CPUs will be affected by the bug quite severely, and will crash regularly even in everyday tasks without the fix installed (but what is the distribution? Is that 1 in every 10 CPUs, 1 in every 100? We may rarely see such a CPU, but even a few of such CPUs could severely hurt AMD's reputation).
I think the message AMD is trying to give off with the disabling of the fix is more like "Try it, see if you can get away with it".

As for AMD obscuring the issue... Yes I think so. Prior to launch, they continually lowered the launch speeds. At first it would be 2.6 GHz, then 2.4 GHz, and a day before launch, even the 2.4 GHz was dropped.
Another issue is that they sent out engineering samples to reviewers with a nb clock of 2 GHz, and production models are only 1.8 GHz.
(And ofcourse the Opteron launching even lower, only at 2 GHz).

I think these two actions from AMD show that AMD knew there was something up with the chips, and thought that they could fix it by reducing the clockspeeds. They probably hoped to get away with it at 2.3 GHz, so they could delay the 2.4 GHz and faster models until they had a proper fix. But no, they could never completely avoid the bug on all CPUs, so they had to go for the BIOS fix. They wanted to avoid this at all costs ofcourse, because it has quite a severe performance impact.
Can't blame AMD for that though, they had a lot riding on this new architecture.
 
Pepsiennis said:
Hey Manny! Well, it's not BS. Tech Report went overboard with their reporting. So much so that Scott Watson even issued an apology in intimating that AMD had actively deceived press and public.
Click to expand...



I stand by my post,it is BS.I have read the apology,hours after it was first posted.

Scotts 'apology' was not a blanket statement,or in any way,shape or form,a complete retraction of his three articles.Far,far from it. Do not try to make it out as though it is,or was. I think price plays a bigger role in the average Chinese citizens choice,when computer parts is concerned,and not overall performance.
 
You must log in or register to reply here.
Back
Top