How did AMD get to 45nm without HKMG?

[RiverRicer]

I was interested in what improvements were made to Deneb/Shanghai in order to get increased performance at lower power.....so a'Googlin I went. Found this link to a pretty decent article that goes beyond the regurgitated tripe found on most 'news' sites.

http://www.eetimes.eu/semi/212002481...RSKH0CJUNN2JVN

Don't let your head explode!

[Glow9]

I was interested in what improvements were made to Deneb/Shanghai in order to get increased performance at lower power.....so a'Googlin I went. Found this link to a pretty decent article that goes beyond the regurgitated tripe found on most 'news' sites.

http://www.eetimes.eu/semi/212002481...RSKH0CJUNN2JVN

Don't let your head explode!

*Shrug* Ask IBM?

[RiverRicer]

*Shrug* Ask IBM?

No doubt, since they are partnered.
I was more interested in what changes were made to the transistor structure.

[Sparky]

No really, what is scary is I actually understood a fair amount of that

[User9498]

No really, what is scary is I actually understood a fair amount of that

Congratulations. You're now one-thousand times more eminently qualified to run your own hardware review site than Kyle Bennett.

[Aussie FX]

DFM has a lot to do with it. AMD decided it wasn't needed at 65nm where intel thought it was. The results speak for themselves, now look at what it has done for 45nm.

I would love to see the FO4 numbers, I think they will have dropped substantially.

User9498 That was priceless.

[Kayin]

AMD wins another one.

To everyone that says that AMD is behind, look at the resources available, and then look at what they produce. Then, if you're not sold, look at their relative sizes.

Besides, articles like this warm the les of my heart.

[dirtyd]

I was interested in what improvements were made to Deneb/Shanghai in order to get increased performance at lower power.....so a'Googlin I went. Found this link to a pretty decent article that goes beyond the regurgitated tripe found on most 'news' sites.

http://www.eetimes.eu/semi/212002481...RSKH0CJUNN2JVN

Don't let your head explode!

I liked this bit:

Providing opinion on who is better when it comes to something as complicated as a multicore microprocessor is a guessing game at best.

[Glow9]

What are all these wins of AMD's that you allude to?

Yeah I love AMD and all but I haven't seen them win a battle in a while. Battles are won by sales and I have to say that chunk they bit out of Intel a few years back was lost. Fortunately in terms of ATI I'd say theres a battle just won recently.

[Vienna]

I tried reading that article, after about the 3rd page my brain was in a meltdown, it just became meaningless words at that point :P

[Apokalipse]

Yeah I love AMD and all but I haven't seen them win a battle in a while.

There's battles of performance, and price/performance.

[mockingbird]

Very interesting.

Thanks for the read.

It made me wonder whether we really need competition in the market. It's obviously evident that most of the research money that has gone towards these engineering processes must have come from income taxes and other taxes of people around the world.

Do we really need this much power now? So that users can have 99% uptime on their mission critical search engine :banana::banana::banana::banana:ography queries? And so that governments can set up huge data harvesting warehouses?

To be honest, I would have been happy with a good optimized dual core Dothan, as we've anyways seen that most optimizations have come from better communication busses, and not pure processing power.

Coupled with that that programmers would have been held responsible for their bloated software (Cough *microsoft* cough *linux too*), more programmers would have been employed, and thus more people would be able to afford adequate technologies for their purposes.

What you have here instead are systems and platforms which will always cost a fortune, with only financial respite coming from the fact that they or their platforms will be manufactured under slave labor.

*clap clap* A huge modern accomplishment for the "evolved man".

/rant

[Nedjo]

what a rant by mockingbird... pfff

[ubuntu83]

bla bla bla

www.apple.com I think you are looking for this.

[Sparky]

Very interesting.

Thanks for the read.

It made me wonder whether we really need competition in the market. It's obviously evident that most of the research money that has gone towards these engineering processes must have come from income taxes and other taxes of people around the world.

*snip*

Coupled with that that programmers would have been held responsible for their bloated software (Cough *microsoft* cough *linux too*), more programmers would have been employed, and thus more people would be able to afford adequate technologies for their purposes.*snip*

You are forgetting something - No competition means the only guy calls the shots on the prices, so no, more people couldn't afford the technologies, rather fewer would That dothan you referred to would very possibly be a premium chip to be had for $500+ if it was a one player field

Competition keeps both parties honest... well, sort of anyway, they can't go too far off the deep end at least.

[mockingbird]

You are forgetting something - No competition means the only guy calls the shots on the prices, so no, more people couldn't afford the technologies, rather fewer would That dothan you referred to would very possibly be a premium chip to be had for $500+ if it was a one player field

Competition keeps both parties honest... well, sort of anyway, they can't go too far off the deep end at least.

I also used to think that but it's not true. I've seen this argument being used before but I've now come to realize that instead of taxes being used to fund the R&D of one monolith it is noe being used to fund the redundant research of two.

Prices of CPUs would have come down regardless of whether there was a competing company because demand would have grown so high due to demand, just as the price for harddrives have fallen because of demand from data warehouses.

[Sparky]

Not really.

Prices would come down eventually yes, but they certainly wouldn't be as cheap as they are now. Why would they, if intel could sell a P4 for $300 because that is all there is, why wouldn't they?

And show me proof that our tax dollars are being used by intel/AMD for research?

*edit* actually don't. Let's not derail the thread please.

[mockingbird]

Not really.

Prices would come down eventually yes, but they certainly wouldn't be as cheap as they are now. Why would they, if intel could sell a P4 for $300 because that is all there is, why wouldn't they?

And show me proof that our tax dollars are being used by intel/AMD for research?

*edit* actually don't. Let's not derail the thread please.

Intel could never charge $300 for a P4 because Willamette was slower than even Katmai never mind Coppermine or Tualatin in its initial release.

Like I said. Breakthroughs came with DDR technology, Dual channel memory, better chipsets, faster GPUs.

[Sparky]

Yes they could have with no competition. I wasn't talking necessarily about initial release.

But like I said, let's not clutter the thread with this off topic stuff.

[demonkevy666]

good work AMD.

[Particle]

Very interesting.

Thanks for the read.

It made me wonder whether we really need competition in the market. It's obviously evident that most of the research money that has gone towards these engineering processes must have come from income taxes and other taxes of people around the world.

Do we really need this much power now? So that users can have 99% uptime on their mission critical search engine :banana::banana::banana::banana:ography queries? And so that governments can set up huge data harvesting warehouses?

To be honest, I would have been happy with a good optimized dual core Dothan, as we've anyways seen that most optimizations have come from better communication busses, and not pure processing power.

Coupled with that that programmers would have been held responsible for their bloated software (Cough *microsoft* cough *linux too*), more programmers would have been employed, and thus more people would be able to afford adequate technologies for their purposes.

What you have here instead are systems and platforms which will always cost a fortune, with only financial respite coming from the fact that they or their platforms will be manufactured under slave labor.

*clap clap* A huge modern accomplishment for the "evolved man".

/rant

You're kidding, right? Nothing in that article suggests the money is coming from the poor like you're making it out. It focuses squarely on the technology. Also, as a programmer and network administrator by trade, I have to say your bloated software conspiracy is pretty weak sauce.

[Rammsteiner]

-quote removed - STEvil

Very interesting.

Thanks for the read.

It made me wonder whether we really need competition in the market. It's obviously evident that most of the research money that has gone towards these engineering processes must have come from income taxes and other taxes of people around the world.

Do we really need this much power now? So that users can have 99% uptime on their mission critical search engine :banana::banana::banana::banana:ography queries? And so that governments can set up huge data harvesting warehouses?

To be honest, I would have been happy with a good optimized dual core Dothan, as we've anyways seen that most optimizations have come from better communication busses, and not pure processing power.

Coupled with that that programmers would have been held responsible for their bloated software (Cough *microsoft* cough *linux too*), more programmers would have been employed, and thus more people would be able to afford adequate technologies for their purposes.

What you have here instead are systems and platforms which will always cost a fortune, with only financial respite coming from the fact that they or their platforms will be manufactured under slave labor.

*clap clap* A huge modern accomplishment for the "evolved man".

/rant

Wut... So is this a rant against taxes? Against pr0n? Against slave labor? Against money? I mean, wake up, we all need those things anyways. Although I can live without pr0n, but for the rest everyone needs that otherwise we wouldnt be anywhere.

Or we can go back living in trees and playing CSS RL

[keithlm]

You know, it's actually tireing me to understand the simplistic nonsense.

Anyway, lets put it this way. They do not win in SuperPI, but their technology is quite good, and with improving their older technology they've done a nice step forward which shows with Deneb.

Sai WUT?

Come ON! If you don't win Super-Pi you might as well pack it up and go buy an e-Machines computer or a Packard Bell.

(Actually the idea of super-pi being important is just as valid as the idea of adding together the percentage losses/gains from a bunch of benchmarks and pretending you have something that actually means something.)

[demonkevy666]

superPI sucks :P

particle's "Generic CPU bench" is Much Better.

[gallag]

SuperPi is used far more nowadays by people on a "All Intel is good for is superPi" rant used in the form of a "Anyway, lets put it this way,They do not win in SuperPI, but their technology is quite good" , "mainly because they all play SuperPi all day long brother LOL" , "We have a launch even today for i7 and I bet they show SuperPi there...as if in the whole scale of things it friggin matters" , "mainly because they all play SuperPi all day long brother LOL (OMFG).....I think my E-PENIS just Grew ....SuPerPi ...FTW" comment.

In fact, I cannot remember the last time I heard someone profess Intel dominance over an AMD cpu based on the results of this bench, Yet it is strangely important to some people that always seem to use it in a derogatory context. Any hobby psychologists wanna take a stab as to why?