ubergizmo
 Ubergizmo Auf Deutsch
Talk Prev Home Next

AMD Tri-Core Drama

Sep 18, 07 11:39 AM PDT

AMD Tri-Core Drama

Ever since AMD has announced its tri-core, most of the drama has been around the fact that it is most likely a defective quad-core, with one disabled unit. It is obvious that it would save AMD money to resell chips that would otherwise go to the trash bin and this might be a hint that AMD's yields (the percentage of functional chips manufactured) aren't that good compared to Intel, but quite frankly, selling partially functional parts it’s not new: Intel did this successfully with the 486 SX (a 486 CPU with the math co-processor disabled). Memory makers also build redundancy in their chips. The end goal is to improve the yields.

Now, should anyone expect AMD to come forward and beg for mercy for selling chips with a disabled core? I don’t think so. Branding it as "The AMD Three-Core Processor. It’s a Quad-Core, only broken" processor would be plain stupid from AMD’s part – nobody wants to buy something with “broken" in it. If they did that, they might as well shoot themselves in the head.

The word “broken” here is confusing at best – the processor is fully functional, simply slower. What matters is the value that consumers get per dollar spent and if a Tri-Core creates a new market segment that is decently priced, it's not bad. There’s no need for additional drama.

( Add a comment)

Source...        Become a fan!

Post a comment

Your email is required but won't be published.

Comments will be published immediately if you verify your post. Be cool, upload your real photo and link your ID to your facebook account.

You are welcome to share your ideas, experiences and questions, but please be respectful of others when commenting. Insulting, self-promotional, SMS-style and off-topic comments will be deleted. Thank you.

blog comments powered by Disqus

Legacy Comments

By KorkiBuchek , 12/04/08 9:26 AM (CommentID #627288)


I am looking at a Gateway GT5670 for $499 w/
AMD Phenom Triple-Core 8400
• 320GB hard drive
• 3GB of DDR2 memory and
NVIDIA GeForce 6150SE vid card.
Not Great system but good buy at $499 ????

Thanks, Korki Buchek

Reply | Direct Link | Report Abuse

By Jake , 01/11/07 6:55 PM (CommentID #350927)


All this talk about AMS's tricore is wasted effort because I am going to buy a Penryn Quad Core in mid 2008...when is the AMD tricore suppose to come out...2009? too late...my money will have gone to Intel by then.

Reply | Direct Link | Report Abuse

By MKK , 27/09/07 8:18 AM (CommentID #314378)


Well, as far as I hear, tri-core clock speed will actually be faster than a quad core because of higher cache per core ratio.

Reply | Direct Link | Report Abuse

By Hubert , 21/09/07 2:53 PM (CommentID #308580)


Slippy, that's a good example. The extra capacity is also some kind of built-in redundancy that leaves room for a few surface defects that are guaranteed to happen, isn't it?

I'm fine with this, as long as I got the GB that I paid for.

Reply | Direct Link | Report Abuse

By Slippy Lane , 21/09/07 1:09 PM (CommentID #308488)


I hate to say it, but you buy something "broken" every time you buy a hard drive that isn't max capacity for the package. If you've got a package that can deliver 80Gig for example, the 40Gig version is the same package but with a defective disk filling the spare space. This has been the way for as long as IBM have made hard drives. I know, I used to build them!

Reply | Direct Link | Report Abuse

By Hubert , 18/09/07 6:09 PM (CommentID #305703)


Great, so I take it that you used to work at Intel back then. Correct?

Reply | Direct Link | Report Abuse

By EPhud , 18/09/07 5:29 PM (CommentID #305670)


You "know" this? How do you know this? Were you there? I was.

Pure urban legend.

Reply | Direct Link | Report Abuse

By Hubert , 18/09/07 3:12 PM (CommentID #305568)


I'm not sure that I agree with that, but I'm interested to know more, of course.

It was a while ado, but I remember that the 486SX was a way to sell 486 CPUs that failed the floating point unit (FPU) test. It is possible that a later design did not feature the FPU at all to save on the die size.

Reply | Direct Link | Report Abuse

By EPhud , 18/09/07 12:54 PM (CommentID #305507)


This claim that Intel sold defective 486DXs as 486SXs is simply not true. It never happened, other than very early samples to get board developers started. It's just one of the urban legends.

Reply | Direct Link | Report Abuse


Recent Reviews


Top Stories


Advertisement

Comment - Report an abuse

Are you sure that you want to report this comment as an abuse?

Report only if you can identify the comment as:

  • SPAM/self promotional
  • Off-topic and unrelated to the current thread
  • Vulgar/offensive/insulting language

Please do not flag a comment as abusive, because you disagree with its content.

Thank you!

Comment - Abuse reported

We will moderate this comment ASAP.

Thank you!

 

  • Andreessen Bullish On Silicon Valley, iPhone
  • Andreessen Calls Google’s Chrome An Extraordinary Event
  • Cisco To Unveil 1st Internal Social Network
  • Google Seeing Docs Use Skyrocket
  • Cloud Computing: New Paradigm Or Technology Repackaged?
  • Pull My Finger App Rejected For iPhone
  • Touch Diamond Almost Released On Sprint
  • Photostream For Android Announced
  • Samsung Omnia Firmware Leaked
  • 5m GBP Fund For Omnia Ads In The UK
  • $39.99 - Crucial 2048MB PC5400 DDR2 667MHz Memory
  • $77.00 - AMD Athlon 64 X2 5400+ Retail
  • $44.99 - A-Data 16Gb USBDrive PD9
  • $86.95 - Linksys WVC54GC Wireless-G Internet Video Camera
  • $372.99 - Acer Aspire One Intel Atom N270 1.6GHz Mini-Notebook