Sign in with
Sign up | Sign in
Your question

Opteron Floating point bug?

Last response: in CPUs
Share
April 28, 2006 12:32:51 PM

http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.
April 28, 2006 2:40:55 PM

If you read the article, there are only 3000 Opterons which might be affected by that bug. And the bug cna be triggered after extreme conditons on high temperature and running a loop that does the same specified intnesive memory operations for more than 4 hours.
So that is far away form the P-bug, that happened to all Pentium chips form that rev.
April 28, 2006 5:55:20 PM

Yup, I did read the article. It says "perhaps 3000". I hope that is the case, I actually have two machines in my lab from that time period and I'd rather not have to go through the hassle of RMAing them.
Related resources
April 29, 2006 9:09:13 AM

It doesn't have anything to do with APM. It's a binning error. The chips do not have a problem, if they are clocked slightly slower.
I'd bet that they were in too big a hurry to get those high priced parts out the door.
April 29, 2006 5:18:16 PM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.


Every IC manufacturer has bugs this isn't something surprising or really problematic in my opinion.
April 29, 2006 5:21:28 PM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.


Every IC manufacturer has bugs this isn't something surprising or really problematic in my opinion.

:)  :)  -- good point, not only bugs within design -- i.e. microcode and logic, but every processor you buy has a defect in it somewhere, some defect are killers, hence the lower cache processors, some defects are annoyances and the die still works....

These die are nusiance defects that turn into failures under certain stress conditions.

Word.
April 29, 2006 6:53:10 PM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.



Ya know what, I have to applaud AMD here, because they discovered the problem admitted it and are actively trying to track down the purchasers of those 3000 CPUs and offering them a free replacement.

Yes, they should have been more careful in the QC phase but 'nuf said.

When is the last time Intel did this?

The Pentium FPU bug was a lot more serious and Intel refused to admit there was even a problem.
April 29, 2006 7:05:00 PM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.



Ya know what, I have to applaud AMD here, because they discovered the problem admitted it and are actively trying to track down the purchasers of those 3000 CPUs and offering them a free replacement.

Yes, they should have been more careful in the QC phase but 'nuf said.

When is the last time Intel did this?

The Pentium FPU bug was a lot more serious and Intel refused to admit there was even a problem.

I remeber things going differently for the Pentium mulitply shifting bug, maybe you have a link to show the evil Intel trying to duck out.
April 29, 2006 7:29:48 PM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.



Ya know what, I have to applaud AMD here, because they discovered the problem admitted it and are actively trying to track down the purchasers of those 3000 CPUs and offering them a free replacement.

Yes, they should have been more careful in the QC phase but 'nuf said.

When is the last time Intel did this?

The Pentium FPU bug was a lot more serious and Intel refused to admit there was even a problem.

I remeber things going differently for the Pentium mulitply shifting bug, maybe you have a link to show the evil Intel trying to duck out.


Quote:


This report stirred up a huge controversy. Intel at first denied that the problem existed. Later, Intel claimed that it was not serious and would not affect most users; however, people who could prove that they were affected would get their processor replaced by Intel. However, although most independent estimates found the bug to be of little importance and have negligible effect on most users, it has caused a great public outcry. Companies like IBM (whose "586" microprocessor competed at that time with the Intel Pentium line) joined the condemnation. Finally, Intel was forced to offer to replace all flawed Pentium processors, at huge potential cost to the company (however, it turned out that only a small fraction of Pentium owners actually bothered to get their chips replaced). As is often the case when a large company gets involved with a public-relations nightmare, Intel's stock price actually rose the day they finally acknowledged what just about everybody else had already realized: that they had to offer a no-strings-attached recall.

April 29, 2006 7:32:33 PM

Aye.

I have to say AMD has behaved a lot better so far.
April 29, 2006 7:35:20 PM

Quote:
I stand corrected, I recall Intel hunkering down for a few days, then confirming the bug....

Good job.


We were in midst of cross posting, just wanted to ensure you note my acknowledgement... again, good job.


Yes, indeed. I just edited my post.

I acknowledge your kind acknowledgement :-D
April 29, 2006 7:43:19 PM

Quote:
Aye.

I have to say AMD has behaved a lot better so far.


I agree, at 3000 CPUs eh, drop in the bucket, on top of that -- I think past history provided a guiding light... as long as they move on this quickly, it will be a simple blimp on headline news and will quickly fade from memory.


I fear some diehard Intel fans will use this to bash AMD.

Given what we know now I do not believe that would be fair.

I agree with you this is a non-issue and I too hope it will quickly fade from memory as it should.

Hiccups happen all the time, every manufacturer has had bad batches. Let's hope they all work to minimize them.

It is the serial offenders that we all need to condemn.
April 29, 2006 7:50:09 PM

Quote:
Aye.

I have to say AMD has behaved a lot better so far.


Better? How? Both companies replaced the faulty products both of which has been described as very minor.
April 29, 2006 7:56:06 PM

Quote:
Aye.

I have to say AMD has behaved a lot better so far.


Better? How? Both companies replaced the faulty products both of which has been described as very minor.


Quote:

Intel at first denied that the problem existed. Later, Intel claimed that it was not serious and would not affect most users; however, people who could prove that they were affected would get their processor replaced by Intel.



Intel aggressively denied the Pentium bug existed at first. Then eventually admitted it under pressure.

In this case AMD has acknowledged the problem and is actively working to track down the purchasers of those CPUs so they can offer them a free replacement.

I have to applaud them for behaving this way.
April 29, 2006 7:59:10 PM

Quote:
Aye.

I have to say AMD has behaved a lot better so far.


Better? How? Both companies replaced the faulty products both of which has been described as very minor.


Quote:

Intel at first denied that the problem existed. Later, Intel claimed that it was not serious and would not affect most users; however, people who could prove that they were affected would get their processor replaced by Intel.



Intel aggressively denied the Pentium bug existed at first. Then eventually admitted it under pressure.

In this case AMD has acknowledged the problem and is actively working to track down the purchasers of those CPUs so they can offer them a free replacement.

I have to applaud them for behaving this way. Yeah because they know what'll happen if they don't.
April 29, 2006 8:44:50 PM

Quote:
Aye.

I have to say AMD has behaved a lot better so far.


Better? How? Both companies replaced the faulty products both of which has been described as very minor.


Quote:

Intel at first denied that the problem existed. Later, Intel claimed that it was not serious and would not affect most users; however, people who could prove that they were affected would get their processor replaced by Intel.



Intel aggressively denied the Pentium bug existed at first. Then eventually admitted it under pressure.

In this case AMD has acknowledged the problem and is actively working to track down the purchasers of those CPUs so they can offer them a free replacement.

I have to applaud them for behaving this way. Yeah because they know what'll happen if they don't.

Word.
May 2, 2006 7:14:51 AM

Quote:
http://www.reghardware.co.uk/2006/04/28/amd_opteron_fpu...

Hmm, I hope this doesn't mean a recall at all. Seems like it doesn't happen that often, though neither did the Pentium bug, and we all know what happened with that one.


Every IC manufacturer has bugs this isn't something surprising or really problematic in my opinion.

Not that I want to pull an older thread forward but I did pull out the errata sheet and it reminded me of your post:

http://www.amd.com/us-en/assets/content_type/white_pape... Look for the section Product Errata

ftp://download.intel.com/design/mobile/SPECUPDT/3092220... Look for section Errata (page 13).

Hmmmmm


I suppose most people are unaware of the existence of these.

Every CPU has issues to varying degrees.
!