AMD Realizes Significant Reduction in Power Consumption by Implementing Cyclos Resona

[grimpr]

AMD Realizes Significant Reduction in Power Consumption by Implementing Cyclos Resonant Clock Mesh Technology

Read more here: http://www.sunherald.com/2012/02/20/3765130/cyclos-semiconductor-announces.html#storylink=cpy

AMD, not ARM, first to use startup's low-power clock IP

Peter Clarke

http://www.eetimes.com/electronics-news/4236592/AMD-counts-on-low-power-clock-IP

2/21/2012 7:32 AM EST

LONDON – Advanced Micro Devices Inc. has achieved the first commercial implementation of resonant clock mesh technology licensed from startup company Cyclos Semiconductor Inc.

AMD’s 4+ GHz x86-64 core code-named “Piledriver” employs resonant clocking to reduce clock distribution power up to 24% while maintaining the low clock-skew target required by high-performance processors. Fabricated in a 32nm CMOS process, Piledriver represents the first volume production-enabled implementation of resonant clock mesh technology. “We were able to seamlessly integrate the Cyclos IP into our existing clock mesh design process so there was no risk to our development schedule,” said Samuel Naffziger, Corporate Fellow at AMD. “Silicon results met our power reduction expectations, we incurred no increase in silicon area, and we were able to use our standard manufacturing process, so the investment and risk in adopting resonant clock mesh technology was well worth it as all of our customers are clamoring for more energy efficient processor designs.”

Read more here: http://www.sunherald.com/2012/02/20/3765130/cyclos-semiconductor-announces.html#storylink=cpy

 

[RavenSEAL]

Now if they could just not screw it up...:hmm:

 

[Rvenger]

Now this is what I want to hear!

 

[gorydetails]

wow,,,reduced power 24 per cent,,,

 

[IlllI]

key words "up to 24%"

then later on in the article it says "The Cyclos inductors and clock control circuits “recycle” the clock power instead of dissipating it on every clock cycle like in a clock tree implementation, which results in a reduction in total IC power consumption of up to 10%.



so i wonder which it is. anyways, i have doubts it would be a 24% decrease across the board 24/7.

​

 

[jones377]

If I understand this right, this 24% power reduction is in the clock distribution network on the CPU, which by itself usually only stands for ~20% of the overall power consumption on a modern CPU design. That would make the overall power savings in the 5% region for the whole CPU. Still not bad though.

 

[grimpr]

Now if they could just not screw it up...:hmm:

I got a feeling that Piledriver chips, Trinity APUs etc..are gonna clock like mad. This tech made AMD make it into Ultrathins with the Trinity 17w APU.

 

[zokudu]

This sounds pretty good. Hopefully Piledriver will be what Bulldozer isn't.

 

[frostedflakes]

A step in the right direction at least, makes me a bit more optimistic about Piledriver, power consumption was definitely one of my biggest gripes with Bulldozer (other stuff like the IPC was disappointing, sure, but wouldn't have bugged me as much if you could clock them to 4.5GHz+ to help make up for the low IPC without having the chips suck down gobs of power).

 

[Raghu]

“We were able to seamlessly integrate the Cyclos IP into our existing clock mesh design process so there was no risk to our development schedule,” said Samuel Naffziger, Corporate Fellow at AMD.

Read more here: http://www.sunherald.com/2012/02/20/3765130/cyclos-semiconductor-announces.html#storylink=cpy

That dude has flip-flops named after him!

Any details on the technique? Journal/research papers?

 

[Khato]

Quite interesting indeed... I'm quite curious to see what kind of overclocking options AMD offers with this change. While the traditional clock tree uses multipliers to obtain the desired clock frequency this is essentially a direct-drive LRC circuit. The capacitance is fixed by the clock mesh while the inductance can be varied in order to obtain the desired frequency, so every change in clock frequency requires selecting a different inductance value. So it becomes a question of how many different values they put into the design.

 

[grimpr]

Quite interesting indeed... I'm quite curious to see what kind of overclocking options AMD offers with this change. While the traditional clock tree uses multipliers to obtain the desired clock frequency this is essentially a direct-drive LRC circuit. The capacitance is fixed by the clock mesh while the inductance can be varied in order to obtain the desired frequency, so every change in clock frequency requires selecting a different inductance value. So it becomes a question of how many different values they put into the design.

Well, they seem to overclock like a piece of cake, according to rumours.

i gotta tell you, no voltage adjustments, no nothing, just cranking the multiplier and whooooosh...the fps gain was nothing short of astounding, and dead stable....AND...that particular unit did NOT have a separated dedicated...just the on apu 7k series.

http://forum.notebookreview.com/8248595-post4815.html

In fact, the trin is so efficient, that the low volt version (17watt) is equiv to our current 3530's, with a better graph card. That is the reason amd decided to trash the brazos replacements, no need, just turn down the voltage on the trinity and rock on. Crank up the voltage and all sorts of good things start to happen Oh, and on these new amd "Ultrathins" , you really have to physically handle the thing to appreciate just how light and thin they are. Astounding. What this remarkable stability suggests, is that the o'clock potential of the new trins is going to be nothing short of outrageous.

http://forum.notebookreview.com/8247215-post4801.html

 

[VirtualLarry]

Wow, this could be really cool. Go AMD! Glad to see further innovation in the market (whether it be from Intel or AMD, or heck Via for that matter.)

 

[pelov]

I really do hope this is true. I've been looking forward to a Trinity APU laptop upgrade over my SB i3 and maybe even offloading my 955 desktop and building a Trinity-powered desktop depending on how well the L3-less chip performs in some desktop apps. I have faith in the AMD APUs and I think they can deliver there. I really don't need anything amazing; I just want something small, efficient and cheap

I don't like rumors but I truly hope I have to eat my words with regard to those high clocked Trinity parts and just what performance they'll be able to bring to the table.

 

[Khato]

Well, they seem to overclock like a piece of cake, according to rumours.

http://forum.notebookreview.com/8248595-post4815.html

http://forum.notebookreview.com/8247215-post4801.html

Well, my initial reaction upon reading those was to laugh, but I guess there's the remote possibility that they're more than fictitious rumor. Especially given that, as stated, if Trinity is indeed using this new clock topology there is no longer a 'multiplier' to crank up. The closest analog would be the circuit inductance, and you'd actually be lowering it in order to increase frequency.

 

[Phynaz]

wow,,,reduced power 24 per cent,,,

Clock power, not cpu power.

 

[grkM3]

Ahh looks like amd is doing what its good at.hyping there next cpu and always keeping people waiting for there next cpu.

This is nothing more than marketing bs to keep the stock holders happy.

 

[busydude]

When is trinity launching?

 

[gmaster456]

Sounds awesome if they can follow through with these claims.

 

[BD231]

10% isn't much of anything in the grand scheme but it'll be interesting to see if this has a positive or negative effect on overclocking, Piledriver should be a more viable option once yields go up to a point where AMD can price it accordingly. Not sure why a company would attest to bowing out of the high end race then charge high end price's, the pricing structure is all messed up right now.

 

[Fox5]

Decent, although this still won't match Ivy Bridge.

I still don't see it as a reason to cancel the Bobcat successors though. Sure, 17W is good enough for the ultrathins (aka netbooks) market, but the bobcat line went even lower power (~5W) and more importantly SHOULD have been super cheap to produce with its die size.

 

[podspi]

Decent, although this still won't match Ivy Bridge.

I still don't see it as a reason to cancel the Bobcat successors though. Sure, 17W is good enough for the ultrathins (aka netbooks) market, but the bobcat line went even lower power (~5W) and more importantly SHOULD have been super cheap to produce with its die size.

Yep. My guess is that AMD looked at the market and realized they could do a light refresh of BC now, and make a go at BC @ 28nm when 28nm is more mature...

 

[Rvenger]

Ill take a power efficient 8 core that has 15% better IPC than bulldozer over a 4 core (i5 based) Ivy. I rather have the multithreaded performance personally because if applications start getting written for more than 4 threads, all of the 4 core Intel CPUs are going to tank in performance.

 

[pm]

It's a really neat idea. Looks like they presented it today at ISSCC 2012. I'll look forward to reading the paper.

 

[IntelUser2000]

Ill take a power efficient 8 core that has 15% better IPC than bulldozer over a 4 core (i5 based) Ivy. I rather have the multithreaded performance personally because if applications start getting written for more than 4 threads, all of the 4 core Intel CPUs are going to tank in performance.

Not really, AMD needed 2x cores vs. the Xeons to be competitive as well. It'll get them in the range of competitiveness, that's it. Piledriver isn't looking bad though.