AMD is already designing products on the 14nm node

[Admire]

Will people ever learn to stop the speculations and hype until we see the results?

This always happens with amd

[Opcode]

Will people ever learn to stop the speculations and hype until we see the results?

This always happens with amd

What is there to hype? All this confirms if true is AMD will be definitely getting some power consumption cuts.

[Admire]

What is there to hype? All this confirms if true is AMD will be definitely getting some power consumption cuts.

 

Sure, but lets wait until we have benchmarks and reviews. Im completely and utterly done with any opinions before we have that. Mantle was also doing to make everything else irrelevant and would catapult amd past nvidia, and a million other things too.

 

I just refuse to wildly speculate anymore....

[Opcode]

Sure, but lets wait until we have benchmarks and reviews. Im completely and utterly done with any opinions before we have that. Mantle was also doing to make everything else irrelevant and would catapult amd past nvidia, and a million other things too.

 

I just refuse to wildly speculate anymore....

It's part of being a true techie. Without wild speculation these forums would be a ghost town. 

[Admire]

It's part of being a true techie. Without wild speculation these forums would be a ghost town. 

 

Not at all. We would just have far fewer topics with not point and far fewer fights. Im sick and tired of this, and people proudly advising people what they should do with their money based on this speculation. "Ohh you should not buy this CPU because this company is coming out with something awesome".....Or something of the same.

[Faa]

I'm only correcting you about what you think you know. Somewhere along your assumptions you dug yourself a hole.

O really, link me where I said a lower manufacturing process (take an existing architecture only shrinked down) will bring performance. 

[Sonefiler]

They better be; isn't Zen on 14nm?

[Opcode]

O really, link me where I said a lower manufacturing process (take an existing architecture only shrinked down) will bring performance. 

Face it, you tried capitalizing on correcting someone who was right to begin with.

 

They better be; isn't Zen on 14nm?

New leaks are suggesting Zen indeed will be 14nm FinFET.

[Faa]

Face it, you tried capitalizing on correcting someone who was right to begin with.

In my first quote to you I've already said that dieshrinking doesn't improve performance, also you've used 14nm & Jim Keller as a reason why AMD will take the performance throne not me.

[Opcode]

In my first quote to you I've already said that dieshrinking doesn't improve performance, also you've used 14nm & Jim Keller as a reason why AMD will take the performance throne not me.

You're the one that's under the assumption that die shrinks are what ultimately decides architectures performance. You can either own up to being wrong or continue on the one way street where you're just dragging out your own demise. No matter how you word it, you were wrong and still are wrong. I would suggest not making assumptions based on your own (often wrong) interpretations of the English language.

[Faa]

You're the one that's under the assumption that die shrinks are what ultimately decides architectures performance. You can either own up to being wrong or continue on the one way street where you're just dragging out your own demise. No matter how you word it, you were wrong and still are wrong. I would suggest not making assumptions based on your own (often wrong) interpretations of the English language.

Sorry but you're the one that used 14nm as an argument why AMD will take the performance throne, you had to correct yourself afterwards. And now stop lying; http://linustechtips.com/main/topic/297577-amd-is-already-designing-products-on-the-14nm-node/page-3#entry4049548

Learn to read

[decripple]

So glad to see some news about AMD on modern lithographic processes. Yeah it might be 2+ years behind Intel, but the R&D still needs to happen IMO.

 

 

[Sonefiler]

Sorry but you're the one that used 14nm as an argument why AMD will take the performance throne, you had to correct yourself afterwards. And now stop lying; http://linustechtips.com/main/topic/297577-amd-is-already-designing-products-on-the-14nm-node/page-3#entry4049548

Learn to read

Dude you're wrong. Give it up. You have no leg to stand on

[TroubleKlef]

Pah! AMD is so behind! We want 1nm nodes! 

[Faa]

Dude you're wrong. Give it up. You have no leg to stand on

I can't be wrong about something I haven't even said. Link me where I said a dieshrink gives more performance. He's the guy after all who said 14nm will make AMD taking the performance throne, he had to correct himself.

[Opcode]

Sorry but you're the one that used 14nm as an argument why AMD will take the performance throne, you had to correct yourself afterwards. And now stop lying; http://linustechtips.com/main/topic/297577-amd-is-already-designing-products-on-the-14nm-node/page-3#entry4049548

Learn to read

Actually both of my posts made prior to this "battle of rightness" capitalizes on die shrinks accordingly.

 

I doubt Intel will spend R&D into developing their own HDL like AMD has. Especially given the limiting factors from using HDL it doesn't exactly fit into Intel's tick/tock cycle. There is a reason why Excavator isn't going to make it to the desktop and it's partial to HDL. Intel is hoping to move away from silicon in 2017 so they have bigger plans in the works than targeting power consumption that they currently are ruling at in the x86 market. I was implying if Carrizo was built on 14nm FinFET it would been a strong competitor to Intel's Broadwell in performance per watt across the board. AMD has a bright future especially with the two massive guns under their belt both Zen and K12 being developed by and in partial with the legendary Jim Keller. Top that with a 14nm FinFET design and even the biggest Intel fanboy has to admit AMD could come back swinging very hard. Everything is lined up in AMD's favor to once again go for the performance throne. I doubt we will see HDL used in the desktop variant of Zen for obvious reasons. Unless AMD can really get their IPC up by that much to where they take us back into the Athlon 64 days where the 2.4 GHz Athlon 64 was beating the snot out of Intel's 3.0+ GHz Pentium 4. Given that exceptional case then there's a possibility that HDL may be used in Zen.

 

What has to be understood is Zen is an entirely new architecture developed from scratch by Jim Keller. With someone who's reputable for developing some of the fastest architectures around while coupled with being on the same node as Intel. AMD has a shot at bringing more performance at similar or less power consumption. Especially factoring in the power improvement technologies that AMD has (resonant clock mesh). Will Zen out perform it's competitor Skylake? We don't know but they do have everything working in their favor to give them the chance to. Even being within a ~10% margin of Skylake performance will be a huge win for AMD.

 

Not sure where your logic is because you say I'm wrong which means you obviously must think one of two things. Either die shrinks only bring performance improvements or die shrinks only bring power improvements in which either alone is wrong. Die shrinks usually do net a performance increase of up to around ~5% in IPC per node. AMD is jumping several nodes from 28nm down to 14nm which should give them up to a ~10% increase in performance without factoring in architecture. The biggest selling point in die shrinks is performance per watt as with smaller nodes the electrical circuits are much smaller thus requiring far less power. I think we all can agree that even a ~10% increase in performance is something AMD could really use at this point even tho it will not ultimately define how Zen will perform. As you need a well thought out and implemented architecture to see real performance. You can't win in this industry even if you have the faster architecture if your microprocessor is eating hypothetically 200 watts to just idle at the desktop.

[Faa]

Not sure where your logic is because you say I'm wrong which means you obviously must think one of two things. Either die shrinks only bring performance improvements or die shrinks only bring power improvements in which either alone is wrong. 

Here you say the shrink has nothing to do with performance and the architecture itself is entirely responsible for performance. You're correcting yourself.

Now it's up to you; proof a die shrink improves performance. Don't try something silly comparing FinFet (Ivy Bridge) with Sandy Bridge, IVB is an exception.

The i5 750 being a 45nm Nehalem and the i5 661 a 32nm, there's just no difference except the 45nm one has 4MB more cache so it will equal out if both used the same amount of a cache. 10% from a 28nm to 14nm is exaggerated unless it's a move to FF. I'd say you won't get any performance gains with a dieshrink only (without the arch being touched), probably at best around ~3%. 

[Opcode]

Here you say the shrink has nothing to do with performance and the architecture itself is entirely responsible for performance. You're correcting yourself.

Now it's up to you; proof a die shrink improves performance. Don't try something silly comparing FinFet (Ivy Bridge) with Sandy Bridge, IVB is an exception.

The i5 750 being a 45nm Nehalem and the i5 661 a 32nm, there's just no difference except the 45nm one has 4MB more cache so it will equal out if both used the same amount of a cache. 10% from a 28nm to 14nm is exaggerated unless it's a move to FF. I'd say you won't get any performance gains with a dieshrink only (without the arch being touched), probably at best around ~3%. 

You're right I did say architecture is solely responsible for performance and I would be exactly right in that respect. If AMD shrunk down Bdver1 do you think they would actually have a chance against Intel? Absolutely not Bdver3 already out performs what would be a 14nm Bdver1 on its existing node. If you're asking me to prove die shrinks improve performance take a look at Intel's tick/tock cycle where they've done nothing to their architecture several times. A prime example is the jump from Haswell to Broadwell which net Intel around a ~4% increase in performance (shorter pipes). It's nothing major but it's still an improvement none the less. Like I said there's no way for you to come out on top of this dispute. You were wrong from the start with your assumptions and are only sinking your ship even deeper the further you drag it out.