The AMD Execution Thread [2007 - 2017]

[Tim Murray]

After reading TR's benchmarks of Phenom with the BIOS fix, I'm a lot more interested in the Linux kernel fix, since that supposedly has a <1% performance penalty (just introduces a few extra page faults).

But then again, that seems to have the disclaimers of "Oh God, if you use this, your machine may explode or open a portal to 2873, where the Insect-Lords of Kron rule Earth with an iron fist."

 

[vertex_shader]

Not much of a statistics but I don't know any better: http://www.steampowered.com/status/survey.html

There entire HD series seem to have around 1.5% marketshare vs ~13% for 8x00 series

Thanks, Valve survey not really informative.
JPR not have any information (or i not find).

 

[nicolasb]

We're in a different situation than HP or Dell because we're the leading provider of fabric computing architecture, which is a disruptor technology

Shields up! Go to red alert!

 

[MTd2]

The problem is not only the TLB bug, but also the low yields. TLB bug is just a lame excuse to justify the abscence in the desktop market. They judge that their reputation must be higher with the high performance maket than with the general public, so they send everything they manage to produce to those costumers.

 

[MTd2]

The Register has received reports of virtualization software slowing by as much as 50 per cent.

We've also confirmed through various channels that it's virtualization code which, in fact, takes the biggest performance hit as a result of AMD's bug cure.

http://www.theregister.co.uk/2007/12/06/amd_benchmarks_down/

 

[Fox5]

The Register has received reports of virtualization software slowing by as much as 50 per cent.

We've also confirmed through various channels that it's virtualization code which, in fact, takes the biggest performance hit as a result of AMD's bug cure.

http://www.theregister.co.uk/2007/12/06/amd_benchmarks_down/

Ouch, goodbye opteron.

 

[BRiT]

Ouch, goodbye opteron.

Double Ouch... The 13-50% lower performance comes with a 14-21% increase in heat generation.

And we round out the woeful news threesome with word that AMD has adjusted the peak power consumption figures on 2.0GHz+ Opterons much higher. Company documents show AMD moving from TDP (thermal design power) levels of 68W, 95W and 120W for various flavors of its chips to 79W, 115W and 137W.

 

[dizietsma]

What would Ned Flanders say at a time like this I wonder ? It would probably include the word "diddly" in it I think.

 

[MTd2]

We wanted to share the last details of the Core 2 Duo, 45nm Wolfdale chips and to reveal their prices. The cheapest of them with 6MB clocked at 2.66MHz with FSB 1333 and LGA 775 packaging, branded as E8200 will sell for as little as $159.

The second in this line E8400 has the same specification but works at 3.0GHz and it will sell for incredibly cheap $179.

The top dual core at 45nm works at 3.16GHz and it will sell for $258. These chips will replace the existing Core 2 Duo Merom based E6x50 series and we are sure that they will overclock well.

http://www.fudzilla.com/index.php?option=com_content&task=view&id=4602&Itemid=35

 

[Bouncing Zabaglione Bros.]

Anyone know what this is about? The implication is that AMD is about to pull a rabbit out of a hat and it's ass out of the fire.

 

[BRiT]

My guesses:

The company was sold... or, an announcement of a dozen or so class action lawsuits...

 

[ShaidarHaran]

Anyone know what this is about? The implication is that AMD is about to pull a rabbit out of a hat and it's ass out of the fire.

Pffft, at this point anything short of an announcement of bankruptcy would be good news.

 

[MTd2]

AMD will announce a will line of 32nm chips to be released this christmas near the North Pole.


Hohoho

 

[AlphaWolf]

AMD will announce a will line of 32nm chips to be released this christmas near the North Pole.

You're close, its an OEM contract with Santa.

 

[INKster]

Anyone know what this is about? The implication is that AMD is about to pull a rabbit out of a hat and it's ass out of the fire.

Depends on who's pulling what out of where.

 

[Geeforcer]

AMD decides that going back to K8 was not far enough, announces K6-IV?

 

[Arun]

I can't really see what this could be except an announcement by AMD that they're now in love with TSMC's 45nm node for CPUs. I don't see how that has anything to do with Barcelona though (Shangai maybe but that's another question!) - maybe they've got Kuma on TSMC's 55nm and thus lol-free capacity for Barcelona? That'd be a very surprising move though, and I'd be curious to see how high it could clock... Or this might actually be related to Fab38, but I question why distributors and SIs should be the first to know about this?

 

[MTd2]

AMD hires Mike Uhler as VP of Accelerated Computing

All that AMD wanted for Christmas was someone to lead its Accelerated Computing division and it seems that Santa made a relative miracle and delivered Mike Uhler. The former CTO (Chief Technology Officer) of MIPS has now been hired by the Barcelona buster and has made him sign up for the position of vice president of Accelerated Computing.

“We are excited to have Mike join AMD just as chip architectures, AMD software development and third-party development of related technologies are converging around our Accelerated Computing vision,” said Phil Hester, senior vice president and chief technology officer for AMD.

AMD’s Accelerated Computing research and development (R&D) programs are focused on speeding up certain tasks and improving a platform's performance-per-watt ratio by putting to good use discrete co-processors and on-chip accelerator cores.

Hopefully Uhler's vast experience will help AMD's division and contribute to getting the whole company on track and working well.


http://www.tcmagazine.com/comments.php?id=17193&catid=6

 

[MTd2]

AMD’s surprise is a new 2GB FireGL


http://www.fudzilla.com/index.php?option=com_content&task=view&id=4637&Itemid=34

 

[MTd2]

Intel is hardcore 32 nm...
Ouch, ouch, damn for AMD.

http://www.x86watch.com/news/intel-silverthorne-samples-184.html