Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

AMD's Piledriver To Hit 4GHz+ With Resonant Clock Mesh

Unknown Lamer posted more than 2 years ago | from the inventing-a-better-sliderule dept.

AMD 286

MojoKid writes about some interesting news from AMD. From the article: "Advanced Micro Devices plans to use resonant clock mesh (PDF) technology developed by Cyclos Semiconductor to push its Piledriver processor architecture to 4GHz and beyond, the company announced at the International Solid State Circuits Conferences (ISSCC) in San Francisco. Cyclos is the only supplier of resonant clock mesh IP, which AMD has licensed and implemented into its x86 Piledriver core for Opteron server processors and Accelerated Processing Units. Resonant clock mesh technology will not only lead to higher clocked processors, but also significant power savings. According to Cyclos, the new technology is capable of reducing power consumption by 10 percent or bumping up clockspeeds by 10 percent without altering the TDP." Unfortunately, aside from a fuzzy whitepaper, actual technical details are all behind IEEE and other paywalls with useless abstracts.

cancel ×

286 comments

Sorry! There are no comments related to the filter you selected.

vaporware (5, Insightful)

networkBoy (774728) | more than 2 years ago | (#39179813)

it's all vaporware till they ship, and it works.
if they pull it off though, might give Intel a run for their money again, it's about time!

Re:vaporware (-1)

Anonymous Coward | more than 2 years ago | (#39179881)

Ah how appropriate! During Black History Month we celebrate another cutting-edge technological innovation invented by African Americans.

Oh wait ... you mean this was invented by whites and Asians? Oh... well then. Business as usual then.

Don't you dare question that contrast either, bub, because that would make you a terrible, horrible person. Not even if you believe it has nothing whatsoever to do with genetics or other inherent superiority (and so fail the definition of a racist not that anybody bothers to learn what the word means) but everything to do with "gangsta" worship and thinking an anti-achievement culture is somehow cool, meaning a change of priorities would turn it around... nope, that still makes you a bad guy!

Re:vaporware (2, Insightful)

Johann Lau (1040920) | more than 2 years ago | (#39181039)

Nah, it just makes you a tool. You're too stupid to be considered a villain, sorry.

Re:vaporware (5, Informative)

Anonymous Coward | more than 2 years ago | (#39179955)

This is an ad. What is a "resonant clock mesh"? That's sounds really cool. So I started RTFA (I know, sorry). You don't have to chastise me that much, because I stopped reading soon. Right after

An average Google search is reported to
require ~ 0.3 watts, about the same amount of power that it takes for a 100 watt light
bulb to be lit for 10 seconds.

Which was obviously not written by anybody who has any clue what they are talking about.

Re:vaporware (2)

TheGoodNamesWereGone (1844118) | more than 2 years ago | (#39180101)

Agreed. It's a breathlessly ebullient press release sales pitch. That said, I hope AMD is able to get back into the game to keep Intel honest, and I own an Intel processor (the last four or five machines I built before it were AMD-based).

Re:vaporware (2, Insightful)

Daniel Phillips (238627) | more than 2 years ago | (#39180281)

The only workstation class machine with which I have been completely happy is powered by an AMD 4 way Phenom II. Quiet, powerful, cheap, pick all three. And looking around, I would say that its successor is highly likely to be an AMD 6 way, 45 nm process chip. Best value by far for my money.

Today I can choose slightly less latency with Intel or significantly more value with AMD. Call me cheap, but I will take the value, thank you.

Re:vaporware (-1)

Anonymous Coward | more than 2 years ago | (#39180515)

If you add up the user-affecting latency over the lifetime of the processor, and multiply by your hourly rate, you would have to be a fairly lowly serf to not pay for it several times over.

Re:vaporware (1, Insightful)

Daniel Phillips (238627) | more than 2 years ago | (#39180815)

If you add up the user-affecting latency over the lifetime of the processor, and multiply by your hourly rate, you would have to be a fairly lowly serf to not pay for it several times over.

Oh indeed, they pay me peanuts. I should send somebody an email.

Or maybe I should just be thankful that today I have a supercomputer[1] class computer under my desk for which I paid peanuts.

[1] As of not very long ago at all.

Re:vaporware (2)

viperidaenz (2515578) | more than 2 years ago | (#39180517)

You first say you picked all three "quiet, "powerful", "cheap". Then you say you dropped the powerful to get the cheap. I'm confused.

Re:vaporware (0)

Anonymous Coward | more than 2 years ago | (#39180773)

"Powerful" doesn't necessarily have to mean "absolutely the most powerful you can get anywhere".

Re:vaporware (0, Flamebait)

Daniel Phillips (238627) | more than 2 years ago | (#39180779)

You first say you picked all three "quiet, "powerful", "cheap". Then you say you dropped the powerful to get the cheap. I'm confused.

Do not be confused, gentle reader. You may understand that as "powerful enough; very powerful indeed". And please do not put words in my mouth.

Re:vaporware (2, Interesting)

Anonymous Coward | more than 2 years ago | (#39180643)

80% of Intel performance at 12% of the cost.

Re:vaporware (1)

billcopc (196330) | more than 2 years ago | (#39180771)

Yep. This is where AMD lives and dies: the budget segment. That's where they stomp Intel, which prefers to keep its high margins and the mindshare that comes along with having the fastest chip of them all.

For myself, AMD would have to push out very affordable 4-socket and 8-socket Opteron solutions, like they did in the K8 days. These days, it's a better value for me to spend the big bucks on Intel workstations and ride them out for an extra year.

And when i say workstation, I'm thinking "server board with GPUs", whereas you seem to be thinking in terms of a standard desktop PC with high quality components. I still think you'd be better served by an Intel i7-2500k, which is also quiet, cheap and powerful (for a single-CPU system).

Re:vaporware (1)

Daniel Phillips (238627) | more than 2 years ago | (#39181087)

These days, it's a better value for me to spend the big bucks on Intel workstations and ride them out for an extra year.

Your strategy confuses me. In the "extra" year you will lose big.

Re:vaporware (2)

haruchai (17472) | more than 2 years ago | (#39181309)

There was a time 8-12 years ago where it looked like AMD could have snatched the performance crown. But, without the Fab expertise to match Chipzilla, it just never happened and nothing short of a fantastic screwup by Intel or an astonishing breakthrough by AMD will close the gap. But, AMD has been rock-solid for my personal needs and make it so easy to keep migrating to newer CPUs / Mainboards that I haven't run an Intel desktop, at home, in 10 years.

Re:vaporware (3, Informative)

hawguy (1600213) | more than 2 years ago | (#39180495)

This is an ad. What is a "resonant clock mesh"? That's sounds really cool. So I started RTFA (I know, sorry). You don't have to chastise me that much, because I stopped reading soon. Right after

An average Google search is reported to
require ~ 0.3 watts, about the same amount of power that it takes for a 100 watt light
bulb to be lit for 10 seconds.

Which was obviously not written by anybody who has any clue what they are talking about.

I think it was a typo (or edit by someone who doesn't know what they are talking about). They should have said 0.3 watt-hours (and should have said "energy" instead of "power")

Google says they use 0.0003 kWh of energy per search [blogspot.com] .

A 100W bulb uses .1 kWh in an hour, or .0000278 kWh in a second, or .000278 kWh in 10 seconds. (or .278 Wh)

Therefore, a 100W bulb running for 10 seconds uses about the same amount as energy as an average Google search. Which is a lot higher than I thought it would be - since I use 20W CFL's, each time I do a google search, that's the equivalent of 50 seconds of light per Google search. Just while typing this reply, I did enough Google searches to light up my room for about 15 minutes.

Re:vaporware (2)

PitaBred (632671) | more than 2 years ago | (#39180557)

I think they take all the routers, networking, cooling, etc. into account as well. Not just the CPU power.

Re:vaporware (1)

elfprince13 (1521333) | more than 2 years ago | (#39180959)

This article [hothardware.com] has an informative diagram.

Re:vaporware (0)

Anonymous Coward | more than 2 years ago | (#39180989)

I think he means. .3 watthours which is accurate for the lightbulb part. That much energy in a fraction of a second probably isn't safe in a computer.

Not really (4, Insightful)

Sycraft-fu (314770) | more than 2 years ago | (#39180093)

Intel is already running at 4GHz+. Ok not officially, but it is almost impossible to find a Sandy Bridge K series that won't easily overclock to 4Ghz or more. I bumped my 2600k to 4GHz. No voltage increase, no messing around, just turned the multiplier up. Zero stability issues, doesn't even draw a ton more power. Basically they are just being conservative for thermal reasons.

The 22nm Ivy Bridge is soon to launch as well. Never mind any potential better OCing, it is faster per clock than SB. Well SB is a good bit faster than Bulldozer (who's architecture Piledriver uses) per clock, sometimes more than a bit (depends on what you are doing).

So no, they'd need way more speed to give Intel any kind of run for their money, unfortunately. What they really need is a better design, something that does better per clock, but of course new designs take a long time and BD itself was quite delayed.

Remember the one and only time AMD did eclipse Intel was during Intel's P4 phase. Intel had decided to go for low work per clock, high clock speed. Well speeds didn't scale as they'd hoped and the P4 was not as powerful for it. AMD chips were tops. However the Core architecture turned all that around. It was very efficient per clock, and each generation just gets better. Meanwhile AMD stagnated on new architectures, and then released Bulldozer which is not that great.

Also they have to fight the losing fab battle. They spun off their fabs and as such aren't investing tons of R&D in it. Well Intel is, and thus are nearly a node ahead of everyone else. Other companies are just in the last few months getting their 32nm node and 28nm half-node production lines rolling out products to retail channels. Intel has their 22nm node process complete and is fabbing chips for retail release in a couple months. So they've got that over AMD, until other fabs catch up, by which time Intel will probably have their 14nm half-node process online in Chandler (the plant construction is in full swing).

Sadly, things are just not good in the x86 competition arena. AMD competes only in a few markets, and Intel seems to edge in more and more. Servers with lots of cores for reasonable prices seems to be the last place they really have an edge, and that is a small market.

I don't want to see a one player game, but AMD has to step it up and this unfortunately is probably not it. If they make it work, expect Intel to just release faster Core i chips with higher TDP specs. The massive OCing success shows they could do so with no problem.

Re:Not really (4, Insightful)

networkBoy (774728) | more than 2 years ago | (#39180245)

I'm a diehard Intel Fanboi. My last AMD was an 80286, I owned an AMD80386DX40, but never used it (acquired it at a swap meet after the P60's had just launched).
Prescott had a use case where it outperformed AMD, but it was very narrow, if your load was highly predictive and did not cause cache misses or branch prediction failures, it owned the AMD. Sadly every workload except straight up numerical number crunching was not so good. I used my 3.6GHz P4 for transscoding video. It was the first machine that I owned where I could encode faster than real time (i.e. movie is 60 min, I could encode in 50).

I really hope this pans out for AMD and brings them a little up into Intel's game. While as you said there has only been one time where AMD flat out bested Intel, there have been several cases where AMD has nailed a particular segment:
* Low cost many cores (data compute clusters).
* Low cost reasonable performance for most end user loads.
* Downright cheap CPU for entry machines.
Every time they've done something they have forced Intel to step up to that segment and improve.
In this case I hope to see not the high spec CPU improvement, but rather the mid-range CPU segment get a very low power option. Somewhere in the i5 equivalent range, but giving desktop performance while sipping mobile levels of power.
It would make building a poor man's compute cluster more feasible from a power and cooling standpoint.
-nB

Re:Not really (3, Insightful)

Intropy (2009018) | more than 2 years ago | (#39180445)

It's not true that the AMD lead was that short. The Athlon came out and was immediately on par with or better than Intel's Pentium IIIs. By the time it was thunderbird vs coppermine/tualatin the lead was pretty sizable. That lasted throughout the Athlon64/Pentium 4 period and into the Core's run until the Core 2 duos arrived. The gap was close for a while with Inte's multi-core processors generally superior, but as little as about a year and a half ago, AMD had the better offering in the X3 than Intel's Core i3. Competition is tight, which has been good for the rest of us.

Re:Not really (1)

PitaBred (632671) | more than 2 years ago | (#39180565)

You can probably go higher. I've got a 2500K that's running 4.8GHz on stock voltages. Basically all K series chips can reliably hit 4.5GHz on stock voltages with adequate cooling.

Re:Not really (1)

lightknight (213164) | more than 2 years ago | (#39180627)

And, as always, so can many of AMD's latest offerings (exceed 4Ghz).

Re:Not really (1)

Nemyst (1383049) | more than 2 years ago | (#39180717)

It's even more ridiculous than that. My motherboard automatically overclocked my 2500K to 4.3GHz. From what I can tell, that 1GHz increase over the stock value isn't even pushing it (temperatures are still ridiculously low, with a 7-Zip benchmark hitting 55C). Granted, aftermarket coolers probably help, but I believe a 0.5-0.75GHz bump on a stock cooler is entirely reasonable.

I have a feeling that Intel might actually be downplaying their default clocks; even under the most terrible conditions, I can't see a 2500K not hitting at least 3.5-3.6GHz. In many ways Intel needs AMD to exist, so perhaps they're limiting the stock clock (that the majority of people will end up using) to give them a running chance.

It's unfortunate, but considering the relative sizes of the two rivals, Intel could easily crush AMD if they so desired. They have the R&D advantage, the fab advantage and the pressure advantage.

Re:Not really (0)

Anonymous Coward | more than 2 years ago | (#39181163)

Intel already shipped 4GHz+ processors. Xeon X5698 dual-core 4.4 GHz, for high-frequency trading systems.

Re:vaporware (1)

ackthpt (218170) | more than 2 years ago | (#39180137)

it's all vaporware till they ship, and it works.
if they pull it off though, might give Intel a run for their money again, it's about time!

Intel is pretty good at catching up, even after Intel said nobody needed 64 bit processors and nobody needed multi core processors, they're right there on top.

Re:vaporware (0)

Anonymous Coward | more than 2 years ago | (#39180623)

even after Intel said nobody needed 64 bit processors and nobody needed multi core processors, they're right there on top

Ever heard of disinformation?

Re:vaporware (1)

AbRASiON (589899) | more than 2 years ago | (#39180207)

"might give Intel a run for their money"
I'm sorry to inform you but you're a little (lot) out of the loop on the current state of Intel and AMD processors available. I'm sure most people here don't want to hear this but the little guy is well and truly down on the ground being kicked in the stomach.

I wouldn't be surprised if one of these CPU's at 5ghz would barely compete with Intels current top shelf items, let alone 4ghz.

Re:vaporware (1, Interesting)

networkBoy (774728) | more than 2 years ago | (#39180301)

Is AMD really doing that badly?
Seriously I am out of the loop from an AMD perspective*, but I assumed they were still rocking the cost/performance on the low end of the CPU ranges, and was hoping this would allow them to push into the mid-range i5 territory.
-nB

*all I work on at work & at home is Intel stuff, so I don't have any relevant AMD info.

Re:vaporware (1, Informative)

the linux geek (799780) | more than 2 years ago | (#39180331)

Bulldozer - their current architecture - was really bad. Slow, mediocre price/performance ratio, and power-hungry. It remains to be seen if Piledriver can make it all better.

Re:vaporware (4, Informative)

ifiwereasculptor (1870574) | more than 2 years ago | (#39180413)

Well, here's AMD on a nutshell:

Brazos, the ultra low power processor, is a success.

Llano, the A series, is actually a very solid product. For the cost of an i3, you get a quad core that is about 1/4 slower overall, but whose integrated graphics about 3 times faster. Actually selling very well.

Bulldozer is a disaster unless all you do is video encoding.

Now, here's the puzzling part: they want to use bulldozer, the failure, as the new core for the A series, the success. I hope they find a way to fix it, otherwise my next rig will have an Intel for the first time in ten years.

Re:vaporware (4, Interesting)

tyrione (134248) | more than 2 years ago | (#39180807)

You must not work in Parallel Programming, doing any heavy engineering analysis/modeling. Taking advantage of all those threads and cores within Bulldozer and utilizing it with OpenCL along with the GPGPUs is a dream come true. More and more modeling environments are leveraging all that this architecture offers, but to you if your game doesn't presently use it it's worthless. To each their own.

Re:vaporware (5, Interesting)

Anthony Mouse (1927662) | more than 2 years ago | (#39181341)

Now, here's the puzzling part: they want to use bulldozer, the failure, as the new core for the A series, the success. I hope they find a way to fix it, otherwise my next rig will have an Intel for the first time in ten years.

I think the people calling bulldozer a failure have the wrong expectations. The core used in the existing A series is a direct descendant of the original Athlon from 1999, which itself was very similar to (and designed by the same people as) the DEC Alpha introduced in 1992, predating even the Pentium Pro. Suffice it to say that there isn't a lot of optimizing left to be done on the design.

Bulldozer is a clean slate. The current implementation has some obvious shortcomings, not least of which that the cache architecture is lame. (The L1 is too small and the L2 latency is too high. They might actually do pretty well to make a smaller, lower latency, non-exclusive L2 and use the extra transistors for a bigger L3 or even an L4.) But that's not a bad thing. It's something they can fix and make future generations faster than the current generation. Which is the problem with the old K10 -- there are no easy little changes left to be made to make it substantially faster than it is now.

The other part of the problem is that people want Bulldozer to be something it's not. It isn't designed for first in class single thread performance. It's designed to have adequate single thread performance while reducing the number of transistors per core so that you can have a lot of cores. It's designed for the server market, in other words. And to a lesser extent the workstation market. They designed something that would let them compete in the space that has the highest margins. So now all the high-end gamers who only care about single thread performance are howling at the moon because AMD concluded it couldn't compete with Intel in that sector and stopped trying.

What you have to realize is that it isn't that the design is flawed. It's that you aren't the target market. They could have built something that achieved 90-100% of Intel's best on single threads instead of 60-80% by doubling the number of transistors per thread and halving the number of threads and cores, but think about who would buy that. PC enthusiasts who comprise about 0% of the market. It wouldn't sell in the server market because the performance per core * number of cores would be lower. It wouldn't sell in the budget market because it would require too many transistors per thread and therefore cost too much to manufacture.

Instead, with Bulldozer they can use more modules and sell to the server market or anyone else with threaded software and then and use fewer modules in combination with a GPU and sell to the budget market and the midrange gaming market, and leave the six dozen howling high-end PC gamers to Intel.

Re:vaporware (1)

AbRASiON (589899) | more than 2 years ago | (#39180435)

Yes, they are doing that badly. The bulldozer was a giant dissapointment. They have nothing on the table for the desktop crowd. At almost all price points it's silly to buy AMD at this time unfortunately. Especially for heat / power usage etc.

Re:vaporware (1)

lightknight (213164) | more than 2 years ago | (#39180689)

*shrugs*

AMD's strategy was to switch to milling out 2 cores or so per unit, aka the Bulldozer architecture, and then stitching them together into a processor. I guess it makes the design more compact / easier to fab.

Re:vaporware (1)

moderatorrater (1095745) | more than 2 years ago | (#39180475)

Yes, because everyone knows that having more cycles is the way to win the processor war. That's why the pentium 4 was so dominant.

Re:vaporware (0)

Anonymous Coward | more than 2 years ago | (#39180683)

The P4 would have done a lot better of it didn't spend so many of its superior cycle count stalled waiting for RAMBUS to fetch data from memory and flushing the pipeline due to incorrectly predicted branches. I still prefer AMD but their current products don't work well for me.

That's nice (1)

iamwhoiamtoday (1177507) | more than 2 years ago | (#39179815)

But how will it scale? How many FLOPS can it pull? GHz doesn't mean squat.....

Re:That's nice (3, Insightful)

networkBoy (774728) | more than 2 years ago | (#39179841)

for a single executing thread of a specific bit width GHz means everything.
The trick is can they scale it to multiple cores/threads, while lowering their power to match Intel's performance/Watt at the high end of the compute arena. If they can do that they will once again pull in DC customers.
-nB

Re:That's nice (1)

mhajicek (1582795) | more than 2 years ago | (#39179979)

Single core performance is all that matters when processing a toolpath for CNC machining. I don't care about power consumption, just higher clock speed and fast memory access (large cache).

Re:That's nice (1)

Kaenneth (82978) | more than 2 years ago | (#39180063)

And for Dwarf Fortress.

Re:That's nice (4, Insightful)

Daniel Phillips (238627) | more than 2 years ago | (#39180317)

Single core performance is all that matters when processing a toolpath for CNC machining.

Rubbish. There is no way your CNC machining app will even get close to the minimum latency that a single AMD core is capable of. What you are really saying is that your vendor is slow to get a clue about parallel programming.

 

Re:That's nice (2)

Jeremi (14640) | more than 2 years ago | (#39180687)

What you are really saying is that your vendor is slow to get a clue about parallel programming.

Maybe there are CNC algorithms that aren't easily parallelizable. Or (more likely) they can be paralellized, but the CNC development teams haven't got around to doing that yet. It doesn't really matter which as far as the consumer is concerned -- in either case, they will want a chip that maximizes single-threaded performance. Finger-pointing doesn't help them one bit, but fast CPUs might.

Re:That's nice (1)

Daniel Phillips (238627) | more than 2 years ago | (#39181147)

Maybe there are CNC algorithms that aren't easily parallelizable.

I doubt that, being somewhat familiar with the problem space.

It doesn't really matter which as far as the consumer is concerned -- in either case, they will want a chip that maximizes single-threaded performance.

Speak for yourself. I prefer to keep the money in my pocket, and spend it on more frequent full-box upgrades. This keeps me ahead of the curve on average. Example: in a past gig where money was no object I started life with a Core2 class desktop which was state of the art at the time, but no, even when money is no object the beancounters will reject the idea of a new box every six months. In short order my onetime shiny Intel box was being smoked by your bog standard mail order AMD box.

Or (more likely) they can be paralellized, but the CNC development teams haven't got around to doing that yet.

Yes: slow to get a clue.

Re:That's nice (3, Informative)

Shark (78448) | more than 2 years ago | (#39180803)

He's not talking about running the g-code, he's talking about generating it from a model. Most CAM software are very CPU intensive for toolpath generation.

Re:That's nice (1)

Daniel Phillips (238627) | more than 2 years ago | (#39181171)

Most CAM software are very CPU intensive for toolpath generation.

All the more reason to parallelize it, cutting latency drastically in the process.

Re:That's nice (1)

DJRumpy (1345787) | more than 2 years ago | (#39179991)

Agree. The multi-core trend was more to address inefficiency in CPU design, as well as technological limitations in clock speed. In short, GHz is important, as long as it's efficient.

Re:That's nice (1)

Daniel Phillips (238627) | more than 2 years ago | (#39180551)

Agree. The multi-core trend was more to address inefficiency in CPU design, as well as technological limitations in clock speed.

More precisely, it is about seeking the best tradeoff in the Latency*Heat*Cost equation.

In short, GHz is important, as long as it's efficient.

Interesting proposition. I think its a little more complex than that. The main use of GHz today is to paper over the inefficiencies of current-generation single threaded software.

Re:That's nice (0)

Anonymous Coward | more than 2 years ago | (#39180045)

Is branch prediction/cache misses no longer an issue?
Is cycles per instruction no longer an issue?
If yes to both, is it because high-end processors are pretty much on par with those issues?

It's been awhile since I knew anything about anything, but I assumed that there's still a market for processors whose throughput depends more on other things than on clock speed.

Re:That's nice (1)

networkBoy (774728) | more than 2 years ago | (#39180129)

sure they are still issues, but in addressing the GGP of GHz not meaning squat, GHz still matters. If cost/power/threadcount/and per clock average performance are all the same which would you rather have? 3.67GHx or 4Ghz (~10% higher)?
-nB

Even for single thread it doesn't (1)

Sycraft-fu (314770) | more than 2 years ago | (#39180159)

See Athlon vs P4. Both were best for single threaded stuff, owing to a single core. However the Athlon did more with less, got better performance at lower clocks. Why? It could do more per clock, or more properly took less clocks to execute an instruction.

IPC matters and the Core i series is really good at it. Bulldozer, not as good. What that means is that all other things being equal, BD needs to be clocked higher than SB to do the same calculations in the same time.

Well that is also a problem because the Core i series are beasts with regards to clock speed. You more or less cannot find a k series part that won't overclock to 4GHz on stock cooling at stock voltages with no stability issues.

You need both good IPC and good clock speed for bitchin' single threaded performance. Really the only thing Bulldozer has going for it is that it isn't a true 4 core system, in the classic way of thinking about it. It isn't a full 8 cores, but it is more than just 4 cores with 2 threads per core. So that can help for highly parallel stuff. Unfortunately, usually even in those cases SB wins out, and there is plenty that is not so highly parallel.

Re:Even for single thread it doesn't (1)

PitaBred (632671) | more than 2 years ago | (#39180617)

If you go aftermarket cooling, you can almost certainly hit 4.5GHz on stock voltages. Right now I'm doing 4.8GHz on stock voltage with a 2500K.

Re:That's nice (1)

aepurniet (995777) | more than 2 years ago | (#39180261)

for a single executing thread of a specific bit width GHz means everything.

if the pipeline length of a processor is 2x as long, then it will be equivalent in performance to a processor 1/2 the speed. combine this with the fact that different instructions have different pipeline lengths, and you have a recipe for the GHZ to tell you squat. would you rather have a 2008 netburst pentium 3 or a brand new intel chip? both were available clocked at 3ghz.

Re:That's nice (1)

viperidaenz (2515578) | more than 2 years ago | (#39180637)

Who told you that? It may be true if every n/2 instruction is a branch (where n = pipeline length).

You'll probably find netburst was p4, not p3.

Re:That's nice (0)

Anonymous Coward | more than 2 years ago | (#39180955)

my processor design class at an elite east coast ivy league school. clock speed just determines the length of a pipeline stage. if a processor has twice the pipeline length, it also needs twice the speed. this is why instructions per clock are also an important metric. unfortunately for AMD Intel has been winning both battles.

Re:That's nice (0)

Anonymous Coward | more than 2 years ago | (#39180319)

for a single executing thread of a specific bit width GHz means everything.
The trick is can they scale it to multiple cores/threads, while lowering their power to match Intel's performance/Watt at the high end of the compute arena. If they can do that they will once again pull in DC customers.
-nB

Without knowing CPI, the clock means nothing. This isn't nitpicking, CPI varies GREATLY.
It's like talking about an engine's RPM without knowing torque.

4GHz is like your car doing 10k RPM
Interesting... but useless by itself.

clock mesh? (-1)

Anonymous Coward | more than 2 years ago | (#39179829)

I'll bet this resonant clock mesh gets clogged with dust blown by the fan, and it will overheat and melt.

Misread (1)

Anonymous Coward | more than 2 years ago | (#39179867)

Misread the title as "AMD's Piledriver To Hit 4GHz+ With Resonant Cloak Mesh." Must say, thats a lot cooler than the reality.

Support FRPAA if you're sick of paywalls (-1)

Anonymous Coward | more than 2 years ago | (#39179905)

I know it won't apply in that case because it is privately funded research but how many times did you hit a paywall that prevented you to read papers funded by your tax money. Please take the time to sign the white house petition on american taxpayers having access to research paid by american taxpayers
https://wwws.whitehouse.gov/petitions/!/petition/strengthen-public-access-publicly-funded-research-and-support-federal-research-public-access-act/jF4mxRc4

but really (0)

arbiter1 (1204146) | more than 2 years ago | (#39179929)

Less it can provide competition for Intel's cpu's at same price level and not use a ton more power like they have can't say its any point to care.

Re:but really (2, Insightful)

Anonymous Coward | more than 2 years ago | (#39179975)

Man, I had to read that 4 times and I'm still not quite exactly sure what you're saying.

Let me give it a stab.

Unless it can provide competition for Intel's CPUs at the same price level, and not use a ton more power to do it (as they have been doing recently), I don't think there is any point in caring.

Communication isn't just about belching words, but actually putting them down so people can understand them.

Re:but really (3, Interesting)

arbiter1 (1204146) | more than 2 years ago | (#39180295)

The bulldozer and i7-2600k were about same performance wise but that is 8 core cpu vs 4 cores + HT. Powerusage of both machines at wall was like 250watts under load. When you overclocked both the bulldozer to 4.8ghz and i7 to 5ghz, i7 used 80 more watts, the bulldozer doubled its draw to over 500 watts, i think it was 550 watts.

details? (5, Insightful)

rudy_wayne (414635) | more than 2 years ago | (#39180021)

Unfortunately, aside from a fuzzy whitepaper, actual technical details are all behind IEEE and other paywalls with useless abstracts.

So why post an article that contains no meaningful information?

Oh wait . . . never mind. I forgot where I was.

There are no technical details. (4, Funny)

game kid (805301) | more than 2 years ago | (#39180185)

Cyclos is the only supplier of resonant clock mesh IP.

There are no technical details. It's intellectual property, so it's powered by pixie dust, mana potions, and lawyers. Can't get more meaningful than that.

Re:There are no technical details. (1)

wbr1 (2538558) | more than 2 years ago | (#39180473)

You forgot unicorn farts. Fail. :)

Re:There are no technical details. (0)

Anonymous Coward | more than 2 years ago | (#39180559)

It's intellectual property, so it's powered by pixie dust, mana potions, and lawyers.

You forgot unicorn farts. Fail. :)

He did not. He said layers.

Re:There are no technical details. (0)

Anonymous Coward | more than 2 years ago | (#39181273)

You forgot wit. Fail. :(

ghz after amd says clock speed irrelevant (0)

Anonymous Coward | more than 2 years ago | (#39180049)

Whatever makes a better processor is a good thing, but I find it ironic AMD promoting higher clock speeds after renaming their processors due to the clock speed wars.

Re:ghz after amd says clock speed irrelevant (1)

Daniel Phillips (238627) | more than 2 years ago | (#39180631)

Whatever makes a better processor is a good thing, but I find it ironic AMD promoting higher clock speeds after renaming their processors due to the clock speed wars.

It is not ironic, rather it is because returns from superscalar design are diminishing while feature size keeps shrinking and other incremental technology improvements keep delivering higher practical clock rates.

Resonant Clock Mesh? (0, Funny)

Anonymous Coward | more than 2 years ago | (#39180109)

Sounds fake. Like Hyperthreading. Sounds like they are doing some trick to make the numbers better, but not really improve any performance.

But I'm just an armchair slashfag know-it-all.

Re:Resonant Clock Mesh? (1)

MobileTatsu-NJG (946591) | more than 2 years ago | (#39180621)

But I'm just an armchair slashfag know-it-all.

Hah. If I were a bettin man I'd say you caught the recent story about CERN and the loose cable.

Real Stuff (0)

Anonymous Coward | more than 2 years ago | (#39180201)

This is a real thing. I'm sure anyone who has been reading the research literature knows that it works, but it's just very difficult to do well.

IBM POWER 6 (1)

Anonymous Coward | more than 2 years ago | (#39180275)

IBM was selling POWER 6 processors running at 5Ghz years ago.

Re:IBM POWER 6 (1)

the linux geek (799780) | more than 2 years ago | (#39180355)

And they still sell Power7 with 8 cores and issuing 6 instructions per cycle at 4GHz+. They're obscenely fast, but they're also not cheap unless you're comparing them to Itanium, SPARC, or Intel's -EX series Xeons.

OMG an Intel killer! (-1)

Anonymous Coward | more than 2 years ago | (#39180299)

As we know, whenever AMD says they have an offering that outperforms Intel, they never release anything that underwhelms.

I'm so grateful they announced this. As we know, everyone is buying processors for gaming machines and servers. For a minute I was afraid they'd buy into the hype that they should give up the high end and focus on $379 laptops, ultra-lights, and mobile computing.

I'm going to pour my entire savings into AMD stock - anyone else with me?

awesome (3, Insightful)

smash (1351) | more than 2 years ago | (#39180347)

Maybe it will catch up to the Sandy Bridge Core i5 now?

Synopsys (1)

Anonymous Coward | more than 2 years ago | (#39180359)

-Each logic gate in a chip needs a clock signal to get to it.

-This is normally done via a wavy, wormy mesh of clock wires.

-Clock skew (When clock is sent, it takes x microseconds to traverse the chip) scales exponentially. 10% at 1ghz is 100mhz skew; 10% at 3ghz is 300mhz skew. And so on.

-Clock skew = VERY BAD, big limiting factor in making faster chips.

-Cyloes has solution.

A: Replace the clock with a simple "Tank Circuit" clock, to reduce the [possibility of the clock not working.

B: Replace a massive mesh of interweaving wires with a "clock plane". Most PCB's have a voltage and grounding plane, why not a clock plane?

This design principal has some advantages:

+Less length of wire in the processor = more savings. As hz increases, the savings in power are exponential. As you increase hz the entire name of the game becomes figuring out how to put fewer electrons through the die. Less wiring = less resistance = less power = less heat = more potential for speed.

+No more designing parts of the processor to time around each other; the whole plate loads and unloads very predictably now and superconductive materials can be inserted between the clock and plate to increase saturation speed.

Realworld:
10-35% decrease in power usage.
Scaling the processor to more transistors or more ghz is now much less problematic.

Re:Synopsys (1)

Thiez (1281866) | more than 2 years ago | (#39180745)

> Clock skew (When clock is sent, it takes x microseconds to traverse the chip) scales exponentially. 10% at 1ghz is 100mhz skew; 10% at 3ghz is 300mhz skew.

So when you triple the clockspeed, you triple the skew? That sounds... linear, and not exponential at all.

Sorry AMD (0)

Osgeld (1900440) | more than 2 years ago | (#39180377)

I am currently currently specing out my next two new machines, and for the first time since 1999 I am going intel. The X2 and X3 that are being replaced were just a pain in the ass, and directly out of the box felt unimpressive.

And its not like they were a great deal or anything, less than 20$ difference, so whats the angle?

Re:Sorry AMD (0)

EmagGeek (574360) | more than 2 years ago | (#39180421)

Enjoy all those legacy SATA2 ports that won't support modern SSDs or hard disks.

Re:Sorry AMD (1)

Osgeld (1900440) | more than 2 years ago | (#39180635)

yea cause its a huge problem now with my SATA2 ports and no SSD's cause it really doesn't bother me to wait an extra 3 seconds for the thing to boot once or twice a month

I need more brain power, and could not care less about loading screens

Re:Sorry AMD (0)

Anonymous Coward | more than 2 years ago | (#39180649)

>> I need more brain power

Well that much is certain.

Re:Sorry AMD (0)

Anonymous Coward | more than 2 years ago | (#39181055)

"yea cause its a huge problem now with my SATA2 ports and no SSD's"

While SATA3.0 would be nice, even my 550MB/s(80k IO) read 400MB/s(36k IO) write SSD rarely breaks 60MB/sec except when doing heavy multi-tasking. Doing benchmarks of sequential or random IO both pegs my SATA2.0 port, but those are synthetic.

My guess is that any single given app that only reads a single file at a time, typically only has 1 outstanding IO at any one moment. One would have to write code differently to thread IO requests properly. Like reading from multiple files at a given time and possibly making multiple sequential IOs at the same time instead of waiting for the last one to come back.

Re:Sorry AMD (1)

PitaBred (632671) | more than 2 years ago | (#39180677)

They don't? Since when? There's not a single hard disk out there that will saturate an SATA2 port. And all modern Intel boards have at least 2 SATA3 ports.

Re:Sorry AMD (1)

Anonymous Coward | more than 2 years ago | (#39181091)

The change from SATA2 to SATA3 involved a whole lot more than just increasing the data rate.

The 3.0 specification contains the following changes:

        6 Gbit/s for scalable performance
        Continued compatibility with SAS, including SAS 6 Gbit/s. "A SAS domain may support attachment to and control of unmodified SATA devices connected directly into the SAS domain using the Serial ATA Tunneled Protocol (STP)" from the SATA_Revision_3_0_Gold specification.
        Isochronous Native Command Queuing (NCQ) streaming command to enable isochronous quality of service data transfers for streaming digital content applications.
        An NCQ Management feature that helps optimize performance by enabling host processing and management of outstanding NCQ commands.
        Improved power management capabilities.
        A small low insertion force (LIF) connector for more compact 1.8-inch storage devices.
        A connector designed to accommodate 7 mm optical disk drives for thinner and lighter notebooks.
        Alignment with the INCITS ATA8-ACS standard.

Re:Sorry AMD (1)

viperidaenz (2515578) | more than 2 years ago | (#39180731)

Thats why the intel Q67 chipset has 6 sata3 ports?

Re:Sorry AMD (1)

EmagGeek (574360) | more than 2 years ago | (#39181113)

Um, no, it doesn't. It supports two SATA3 ports and four SATA2 ports.

amd boards have better pci-e I/o and lanes (0)

Joe_Dragon (2206452) | more than 2 years ago | (#39180447)

amd boards have better pci-e I/o and lanes then Intel boards

Re:Sorry AMD (0)

Anonymous Coward | more than 2 years ago | (#39180587)

Really I have the 720 x3 and have had no problems with it. Also it overclocks great i can get 3.3ghz stable with no problems.

Re:Sorry AMD (1)

Osgeld (1900440) | more than 2 years ago | (#39180657)

I have no problem with my720, I never stated it was unstable and crashing

Really? (1)

stazeii (1148459) | more than 2 years ago | (#39180459)

So, are they going to give their chips porn names? Piledriver? Someone needs to clue these guys in... Just calling your stuff by construction equipment names doesn't make it cool. I'm a Mac user, and I still hate the fact Apple has latched onto this "let's call our products by their code names" crap. Guess I should look forward to the days of AMD's Cleveland Steamer processors.

Maybe its their choice of marketing words (2)

viperidaenz (2515578) | more than 2 years ago | (#39180681)

but clock resonance sounds like it wouldn't play well with changing the clock frequency.

Re:Maybe its their choice of marketing words (0)

Anonymous Coward | more than 2 years ago | (#39180791)

All clocks are based on some fundamental resonant frequency. Their output can be skewed a bit with the use of inductors and capacitor. (Not to mention the harmonic frequencies.)

Re:Maybe its their choice of marketing words (0)

Anonymous Coward | more than 2 years ago | (#39181127)

Wow, I've never seen anyone so confidentially fail basic linear systems theory.

resonate clock mesh (5, Informative)

slew (2918) | more than 2 years ago | (#39180763)

Quick background: Currently clocks on most generic chips today are structured as trees. As you can imagine the fan-out of the clock trees is pretty large and thus require clock buffers/driver circuits which need to be balanced so that clock signal gets to the leaves at about the same time (in a typical design where you don't use a lot of physical design tricks). To ease balancing the propagation delay, the clock tree is often physically looks like a fractalized "H" (just imagine the root clock driving in the center of the crossbar out towards the leaves at the corners of the "H", the wire lengths of the clock tree segments are the same, then the corners the big H driving the center of a smaller "H", etc, etc). Of course at the leaves, there can be some residual imbalance due to small manufacturing variations and wire loading and that has to be accounted for in closing the timing for the chip (to avoid short paths), and ultimatly these imbalances limit the upper frequencies achievable by the chip.

Additional background: In any electrical circuit, there are some so-called resonant frequencies because of the distributed (or lumped) inductance and capacitances in the network. That is some frequencies experience a lot less energy loss than average (for the car analogy buffs, you can get your car to "bounce" quite easily if you bounce it at it's resonant frequency).

The basic idea of the Cyclos technology is to "short-circuit" the middle of the clock tree on the chip with a mesh to make sure all the middle of the clock tree is coordinated to be the same clock (as oppposed to a typical H tree clock, in every stage the jitter builds up from the root). That way you avoid some of the imbalances the limit the upper frequencies achievable by the chip. The reason I say "short-circuit" is that it really isn't a "short circuit". If you just arbitrarily put in a mesh in the middle of a clock tree, although it would tend to get the clocks aligned, it would presents a very large capacitive and inductive load to drive and would likely increase power greatly. **Except** if that mesh was designed so that it resonated at the frequency that you were going to drive the clock, then you can get the benefit of jitter reduction w/o the power cost. Since you get to pick the physical design parameters of the mesh (wire width, length, and grid spacing, and external tank circuit inductance) and the target frequency, theoretically you can design that mesh to be resonant (well, that remains to be seen).

The reason this idea hasn't been used to date is that it's a hard problem to create the mesh with the proper parameters and now the processor really has to just run at that frequency all the time (well, you can do clock cycle eating to approximate lower frequencies). Designers have gotten better at these things now and the area budgets for these types of things have gotten in the affordable range as transistors have gotten smaller.

FWIW, In a pipeline design (like a cpu), sometimes it's advantagous to have a clock-follows-signal clocking topology or even an async strategy instead of a clock tree, but there of course is a complication if there is a loop or cycle in the pipeline (often this happens at say a register file or a bypass path in the pipeline), so that trick is limited in appliciablity, where the mesh idea is really a more general solution to clock network jjitter problems.

Here's a white paper that describes this idea... http://www.cyclos-semi.com/pdfs/time_to_change_the_clocks.pdf [cyclos-semi.com]

Re:resonate clock mesh (3, Interesting)

TheSync (5291) | more than 2 years ago | (#39181023)

How can the mesh be resonant to a square wave (with lots of high frequency harmonics over a huge band)?

I can imagine it being resonant to a single frequency sine wave.

But if the clock mesh is powered by a sine wave, you have to turn it back into a square wave to drive gates, and to do that you have to compare the clock voltage level with some known voltage levels, and there you may have process inaccuracies.

A tuned circuit ... (4, Insightful)

PPH (736903) | more than 2 years ago | (#39180929)

... inside the processor? Sounds like the end of overclocking.

Desktop CPU space is a big yawnfest (2, Insightful)

Gordo_1 (256312) | more than 2 years ago | (#39181125)

and has been for at least 5 years. A theoretical 10% performance boost? Gimme a break. I upgraded from a Core2Duo E6600 @ 2.4GHz to a quad core i5 2600k which runs at an overclocked 4.5GHz on air... Day to day, the new rig delivers a *mostly* perceptible performance advantage, but nothing earth shattering... I give you several recent changes that felt bigger:

1. Moving from hard drive to SSD
2. Moving from a DirectX9 class GPU to a DirectX 11 GPU (at least in games).
3. Move from pre-JIT JS browser engine to a JIT-engined browser.

As far as desktop CPU development goes, I think the future is largely about optimizing software for the multi-core architectures, not adding Gigahertz.

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?