Leaked AMD Ryzen Benchmarks?

[razor1]

You forgot about 3.6 GHZ base and 4.0 Boost for the 1800X??? Convenient loss of memory??? I sincerely hope your not on beta blockers or statins like me.
If you are implying the 1700X and 1800X can not function on air cooling you are dead wrong. A Cooler Master Maker 8 , a Noctua 15D, or a high-end Thermalright Silver Arrow cpu cooler will allow some reasonable overclocking. Personally I am using an Alphacool Eisbaer 360 . I already installed it in the top of a new case I bought, a Thermaltake Core V71. I expect to be able to overclock my Ryzen 1800x to 4.2 or 4.3 GHZ on all 8 cores.

Yeah good luck with that will wait for reviews to see if they can over clock like that. Maybe very few gold samples for the sounds of it from tweak town.

Not implying anything btw, never do, my posts are very straight forward, no BS around the edges, if you haven't noticed.

Don't know how long you are in the hardware game, but tweaktown was one of the main leakers 5 to 10 years back, their info is usual very good.

Beta blockers can do that but I don't need those, don't have any memory issues, although lack of sleep and a few drinks will , ps. you should ask your doctor about calcium channel blockers, they might help with the memory problems.

 

[razor1]

This old?
http://www.hwbattle.com/bbs/board.php?bo_table=cpumbram&wr_id=39712&ckattempt=1

no CB?

looks like the page was taken down.

 

[Magic Hate Ball]

Yeah good luck with that will wait for reviews to see if they can over clock like that. Maybe very few gold samples for the sounds of it from tweak town.

Don't know how long you are in the hardware game, but tweaktown was one of the main leakers 5 to 10 years back, their info is usual very good.

Beta blockers can do that but I don't need those, don't have any memory issues, although lack of sleep and a few drinks will , ps. you should ask your doctor about calcium channel blockers, they might help with the memory problems.

Razor man, are you saying that Ryzen 1800X won't hit 4ghz without water cooling or something?

You might need to sit down.

 

[razor1]

Razor man, are you saying that Ryzen 1800X won't hit 4ghz without water cooling or something?

You might need to sit down.

Tweaktown if yeah want 4.0 base clock, ya need water cooling.

PS I am sitting down and listening to the video again, I posted the video earlier in this thread! Great to know people can't seem to get the info when its under their noses.

 

[cageymaru]

no CB?

looks like the page was taken down.

Grabbed them before they were taken down. WCCFTECH has them also.
http://www.hardocp.com/news/2017/02/15/ryzen_spy_pics

 

[razor1]

ah cool thx!

 

[Magic Hate Ball]

Tweaktown if yeah want 4.0 base clock, ya need water cooling.

PS I am sitting down and listening to the video again, I posted the video earlier in this thread! Great to know people can't seem to get the info when its under their noses.

Joker and the other fella chatting each other?

Anthony (TWG) was stating that his "sources" were stating 4.0ghz to 4.3ghz were AIO ranges.... but did they have ES, did they have QS? No word on that.

AMD has not shied away from high TDP chips in the past (9590?) I don't see how a 4.0ghz boost 95w TDP rated chip from them will require an AIO... when they've released a stock 220w TDP chip in the past.

EDIT: this is also the duo that almost entirely forgot AMD's push for deep learning GPU hardware... yes very aware of the market.

 

[razor1]

Joker and the other fella chatting each other?

Anthony (TWG) was stating that his "sources" were stating 4.0ghz to 4.3ghz were AIO ranges.... but did they have ES, did they have QS? No word on that.

AMD has not shied away from high TDP chips in the past (9590?) I don't see how a 4.0ghz boost 95w TDP rated chip from them will require an AIO... when they've released a stock 220w TDP chip in the past.

Do you know thermodynamics of silicon how voltage, frequency, temperature, wattage correlate with node sizes and architecture with leakage, resistance, quantum tunneling, electromigration? Your statement about high TDP chips in the past, tells me you don't.

PS, AMD is so far behind in deep learning right now, its not even a blip on the radar, they are farther behind to Intel as Intel is to nV. Without getting their software solutions up to par with CUDA, functionality and readiness, they won't have much push in that market at all.

For them not to even think about AMD in conjunction with DL, can't blame them, since the market is pretty much cornered by nV and Intel is having a tough time getting in, even though coding for Phi should be easier and old code should be more portable to new gen's of Phi. This has nothing to do with hardware its all about software. There are threads on this where I talk about this more in depth about why AMD has such a hard time in HPC and will have such a hard time in DL markets.

 

[Magic Hate Ball]

Do you know thermodynamics of silicon how voltage, frequency, temperature, wattage correlate with node sizes and architecture with leakage, resistance, quantum tunneling, electromigration? Your statement about high TDP chips in the past, tells me you don't.

PS, AMD is so far behind in deep learning right now, its not even a blip on the radar, they are farther behind to Intel as Intel is to nV. Without getting their software solutions up to par with CUDA, functionality and readiness, they won't have much push in that market at all.

For them not to even think about AMD in conjunction with DL, can't blame them, since the market is pretty much corned by nV and Intel is having a tough time getting in, even though coding for Phi should be easier and old code should be more portable to new gen's of Phi.

OK fella.

You're throwing words, but specifically, what we've read about Ryzen is very low voltage, very dynamic clock speed control, and a live demonstration where the Ryzen system used less wattage than a 6900k system doing the same task.

So many watts, such high TDP! My poor Noctua cooler can't handle such wattage, it's incredible!

AMD GPUs have had higher raw GFlops than Nvidia chips in the same price bracket for at least two generations now. We're not talking about rendering pipelines, but pure flops. How is the MI25 Instinct product light years behind what Nvidia has demonstrated for processing capability?

From what I can tell, most of this type of software is custom. There isn't a software suite with "CUDA CORE ACCELERATION" that everyone is using....

 

[doz]

Grabbed them before they were taken down. WCCFTECH has them also.
http://www.hardocp.com/news/2017/02/15/ryzen_spy_pics

That picture badge.....

*ENHANCE*

 

[razor1]

OK fella.

You're throwing words, but specifically, what we've read about Ryzen is very low voltage, very dynamic clock speed control, and a live demonstration where the Ryzen system used less wattage than a 6900k system doing the same task.

So many watts, such high TDP! My poor Noctua cooler can't handle such wattage, it's incredible!

AMD GPUs have had higher raw GFlops than Nvidia chips in the same price bracket for at least two generations now. We're not talking about rendering pipelines, but pure flops. How is the MI25 Instinct product light years behind what Nvidia has demonstrated for processing capability?

From what I can tell, most of this type of software is custom. There isn't a software suite with "CUDA CORE ACCELERATION" that everyone is using....

I'm not throwing words, if you think those are just words and not inter related, guess what you don't know anything about how chips are designed and why certain limits are different based on those things, please look at past posts about these things I have done, or look them up on the internet.

use these phrases in this order

1) silicon thermodynamics voltage node

2) silicon resistance electromigration

3) silicon TDP voltage power limits node

You will get a ton of college level EE papers that will talk about all of these things on how they are interrelated, if you aren't good with Thermodynamics to begin with, look it up first, but be wary, silicon's thermodynmics are a bit different than regular wires. I=VR, ohms law, invert the resistance for silicon.

nV's current software features eclipse AMD's features for specific DL libraries, I suggest you look old posts I have made on this. AMD is well aware of what features they are missing, and also well aware why CUDA was adopted rapidly by the DL community.

And I stated this before, this problem is not about hardware capability its about software features and extensions that are not available for AMD hardware which gives nV a specific performance and flexibility advantage on top of that people in the DL community are comfortable with CUDA as their standard programming language.

 

[pendragon1]

if you guys stop arguing with him he stops...

 

[razor1]

if you guys stop arguing with him he stops...

This is not arguing, this is people being oblivious to basic concepts (yet they think they know what they are talking about) and why certain things might happen that Tweak town stated.

 

[pendragon1]

who cares?! and I mean this for everyone. does it really matter if they know what they are talking aboot? everyones arguing aboot something they don't actually have any experience with because they are not out yet and jump from one thing to another just to try and "win" the ikps*. wait two weeks and let the truth come out.

*internet knowledge pissing contest

 

[Schmave]

This old?
http://www.hwbattle.com/bbs/board.php?bo_table=cpumbram&wr_id=39712&ckattempt=1

That link seems to be password protected now...

 

[razor1]

who cares?! and I mean this for everyone. does it really matter if they know what they are talking aboot? everyones arguing aboot something they don't actually have any experience with because they are not out yet and jump from one thing to another just to try and "win" the ikps*. wait two weeks and let the truth come out.

*internet knowledge pissing contest

If someone doesn't know what they are posting why are they posting instead of asking questions? Do you know who does that? What kind of person does that?

is that what is taught in school these days when you don't know act as if you do know and the problem will go away?

You want this forum to become like Anandtech or Semiaccurate? The lunacy and ignorance of the posts on those forums are unfathomable.

This is not a pissing contest, I don't even need to take a piss for this.

 

[pendragon1]

dude we've been through this how many times?! tell them theyre wrong and that they need to go read/research, call it a day. pages and pages and pages of arguing is ruining the forum.

 

[razor1]

dude we've been through this how many times?! tell them theyre wrong and that they need to go read/research, call it a day. pages and pages and pages of arguing is ruining the forum.

That is exactly what I did. Then you came in.

 

[Nobu]

Well, you could just say "you don't know what you're talking about" and ignore whetever else comes forth from the void, or else discuss the minutiae of how all this stuff actually works, but another thread would be a better place for that.

 

[razor1]

Well, you could just say "you don't know what you're talking about" and ignore whetever else comes forth from the void, or else discuss the minutiae of how all this stuff actually works, but another thread would be a better place for that.

I did that too, and suggested looking up those threads lol.

 

[Nobu]

I did that too, and suggested looking up those threads lol.

Sorry, was writing that while five other posts were made.

 

[razor1]

Sorry, was writing that while five other posts were made.

I hear ya, I'm tired of writing up these types of posts, if someone whats to talk about chips, tdp, silicon lottery, frequency, nodes, they should know their shit before they make general assumptions like their old chips can do it, why can't their new architecture on a new node do it or just ask someone here or another board that might know.

 

[NKD]

This is not arguing, this is people being oblivious to basic concepts (yet they think they know what they are talking about) and why certain things might happen that Tweak town stated.

this is not true though. I mean AIO cooler and high end air coolers perform very similarly. Ofcourse I am talking about single radiator here which falls in AIO coolers category. Now how can one say for certain they are talking about dual radiator here? You see what I am saying? Only reason I have a AIO cooler is because its quiet and its decent. I have seen some high end aircoolers keep the temps close to my cooler.

also you keep repeating that top end is 3.4 with boost when its 3.6 with boost at 4.0. I really have a hard time believing it wont sustain atleast 4ghz on all cores with decent cooling.

 

[razor1]

this is not true though. I mean AIO cooler and high end air coolers perform very similarly. Ofcourse I am talking about single radiator here which falls in AIO coolers category. Now how can one say for certain they are talking about dual radiator here? You see what I am saying? Only reason I have a AIO cooler is because its quiet and its decent. I have seen some high end aircoolers keep the temps close to my cooler.

close might not be enough, 5 degree difference is 5 extra watts at least, and if that changes leakage levels where you need voltage changes to maintain signal cohesion, it becomes 5 watts x X^2 voltage change. (I simplified this there is much more to this but for this purpose its enough to get the point across)

This is all theoretical of course because of what Tweak town is saying, but its plausible because of the lower end chips seeming to have their clocks limited too.

This is why CPU's have lengthy validation periods, because small changes like this, affect the motherboard designs and tolerances can affect other components.

 

[NKD]

close might not be enough, 5 degree difference is 5 extra watts at least, and if that changes leakage levels where you need voltage changes to maintain signal cohesion, it becomes 5 watts x X^2 voltage change. (I simplified this there is much more to this but for this purpose its enough to get the point across)

This is all theoretical of course because of what Tweak town is saying, but its plausible because of the lower end chips seeming to have their clocks limited too.

This is why CPU's have lengthy validation periods, because small changes like this, affect the motherboard designs and tolerances can affect other components.

I understand. But an AIO will get worse as the the chip gets hotter. You see what I am saying a high end cooler will creep up close to a high end single radiator unit. Since the liquid inside will get hot and can't be cooled fast enough. This is why just saying with an AIO you hit 4.2 to 4.3 does not mean you can't get anywhere with high end air.

 

[Gideon]

Personally I think 4.2 will be a realistic thing to get out of these chips, a few golden samples might hit 4.5. We only have a few more days until we get to find out anyway.

 

[Magic Hate Ball]

I hear ya, I'm tired of writing up these types of posts, if someone whats to talk about chips, tdp, silicon lottery, frequency, nodes, they should know their shit before they make general assumptions like their old chips can do it, why can't their new architecture on a new node do it or just ask someone here or another board that might know.

Hi Razor.

I have been following chips, TDPs, and the silicon lottery since before the Athlon was released...

Unexpected market disruptions occur, often quickly. Sometimes, despite superior implementations of certain features (Tile Based Rendering in the Kyro II at a fraction of the cost 17 years ago?) technologies don't pan out due to bad timing or bad luck.

Currently, AMD has a customer in Google for their Deep Learning products for 2017 http://www.forbes.com/sites/aaronti...g-machine-learning-in-the-cloud/#3d5983e14181

I don't know if you've heard of this company, but Google was co-founded by a kid who graduated from the same high school I did. That's not important, or a claim to intelligence, just a fun bit of trivia. Just wanted to make sure you realized this small garage-based company was interested in their silicon.

Yes, Nvidia was there first, but ultimately that isn't a guarantee for anything. But, if you remember 3dfx was also a dominating market force at one point.

The funny thing about your statement is that you fall into this logical fallacy constantly in your posts when I read them. "Intel (or Nvidia) has succeeded in the past, they must surely succeed in the future!" "They have the money/contracts, they must surely dominate!"

It's rather entertaining.

 

[Magic Hate Ball]

close might not be enough, 5 degree difference is 5 extra watts at least, and if that changes leakage levels where you need voltage changes to maintain signal cohesion, it becomes 5 watts x X^2 voltage change. (I simplified this there is much more to this but for this purpose its enough to get the point across)

This is all theoretical of course because of what Tweak town is saying, but its plausible because of the lower end chips seeming to have their clocks limited too.

This is why CPU's have lengthy validation periods, because small changes like this, affect the motherboard designs and tolerances can affect other components.

This is an poorly thought out statement.

Intel i5 7400 is a 3.0ghz base clock 3.5ghz boost CPU with a 65watt TDP for a retail of $182. How is AMD's product an inferior showing with identical clock speeds for 2/3 the price if you ignore a weak iGPU that isn't worth $60?

Lower power, same work, same core, similar clocks? How is this a flaming wreck per your words?

 

[bobzdar]

I'm not throwing words, if you think those are just words and not inter related, guess what you don't know anything about how chips are designed and why certain limits are different based on those things, please look at past posts about these things I have done, or look them up on the internet.

use these phrases in this order

1) silicon thermodynamics voltage node

2) silicon resistance electromigration

3) silicon TDP voltage power limits node

You will get a ton of college level EE papers that will talk about all of these things on how they are interrelated, if you aren't good with Thermodynamics to begin with, look it up first, but be wary, silicon's thermodynmics are a bit different than regular wires. I=VR, ohms law, invert the resistance for silicon.

nV's current software features eclipse AMD's features for specific DL libraries, I suggest you look old posts I have made on this. AMD is well aware of what features they are missing, and also well aware why CUDA was adopted rapidly by the DL community.

And I stated this before, this problem is not about hardware capability its about software features and extensions that are not available for AMD hardware which gives nV a specific performance and flexibility advantage on top of that people in the DL community are comfortable with CUDA as their standard programming language.

I kind of agree,you're just throwing terms out there to scare him away. None of that will tell you how well this thing will overclock over its base clock. Obviously they can't hit 220w at 14nm, but Intel chips have shown they can go up to ~130w on a 91w spec quad at 14nm, and we know the core sizes are similar.

Fwiw, I have a degree in microelectronics and wrote a paper on maximum clock speed at a given node for my semiconductor physics 2 class and none of those terms were in my paper. Maybe that's why I only got an 80 on it? The electromagnetic field on the gate insulator was the limiting factor in my paper as exceeding a certain amount would damage it. Electromigration is a longer term reliability problem,not an out of the box clock limiter.

That was theoretical though,you'd have to know the layout of the chip and test thermals throughout the chip layout to see where localized hot spots are that can limit performance,which all of the research papers in the world won't help you with. It's why you get higher clocks with hyper threading turned off on an Intel vs on,even though everything else is exactly the same. Hyper threading isn't causing more electromigration or even higher gate leakage, it's causing more localized heat somewhere on the chip causing slightly lower clocks. You can read all the papers you want and it won't tell you how or if that will affect clocks without testing it as it's architecture related, not node related. Given the ryzen chips are a completely new architecture that nobody has flogged yet,there's no real way to determine it. Maybe they've figured out how to enable hyper threading with no overclocked penalty? Maybe it's worse? Who knows?

 

[Brackle]

you guys have to realize razor1 is anti AMD. No matter how much you prove him wrong he will continue to argue. He is known to get threads closed.

Right now everything is speculation. And no one right now knows the overclocking potential of Ryzen. It can totally suck ass, or hit 5ghz. Kyle will get to the bottom of that just like he did with Kaby lake.

In Kyle we trust!

 

[OrangeKhrush]

you guys have to realize razor1 is anti AMD. No matter how much you prove him wrong he will continue to argue. He is known to get threads closed.

Right now everything is speculation. And no one right now knows the overclocking potential of Ryzen. It can totally suck ass, or hit 5ghz. Kyle will get to the bottom of that just like he did with Kaby lake.

In Kyle we trust!

I think we could make a big list headed by Juanrga, 2 March will be a bad day for him on the internet.

 

[OrangeKhrush]

What we do know:

Likely have Sandy or sub Sandy Instructions per Clock, as shown by PM single thread a 3.3/3.4Ghz 2500/2600K get about 1600 to Ryzen's 2050. There is a very causal link to that so must be Sandy clock vs clock level notwithstanding my 3.4ghz Haswell nets slightly lower scores.

It probably wont make it past 3.5ghz because of stuff.

Prime and Physics are now the benchmark of choice

All point to something like a Nahelem/Sandcheeks.

 

[KazeoHin]

What we do know:

Likely have Sandy or sub Sandy Instructions per Clock, as shown by PM single thread a 3.3/3.4Ghz 2500/2600K get about 1600 to Ryzen's 2050. There is a very causal link to that so must be Sandy clock vs clock level notwithstanding my 3.4ghz Haswell nets slightly lower scores.

It probably wont make it past 3.5ghz because of stuff.

Prime and Physics are now the benchmark of choice

All point to something like a Nahelem/Sandcheeks.

Not really, the IPC has been seen at Haswell levels for some time. Some people like to claim it is better than boradwell, but I would think that this is only in certain tasks. Sandy IPC would be pretty lousy.

Also, the cores can go above 3.5GHz, but I doubt ALL of them can go over that and maintain the 95w TDP. If one core can reach 4.0GHz, then given enough TDP allowence, they all can. Wether these boost clocks are an absolute garentee or a rare sight: it remains to be seen

 

[OrangeKhrush]

Not really, the IPC has been seen at Haswell levels for some time. Some people like to claim it is better than boradwell, but I would think that this is only in certain tasks. Sandy IPC would be pretty lousy.

Also, the cores can go above 3.5GHz, but I doubt ALL of them can go over that and maintain the 95w TDP. If one core can reach 4.0GHz, then given enough TDP allowence, they all can. Wether these boost clocks are an absolute garentee or a rare sight: it remains to be seen

sorry I forgot to put on my sarcasm text

 

[KazeoHin]

Poe's Law, my friend.

 

[CSI_PC]

this is not true though. I mean AIO cooler and high end air coolers perform very similarly. Ofcourse I am talking about single radiator here which falls in AIO coolers category. Now how can one say for certain they are talking about dual radiator here? You see what I am saying? Only reason I have a AIO cooler is because its quiet and its decent. I have seen some high end aircoolers keep the temps close to my cooler.

also you keep repeating that top end is 3.4 with boost when its 3.6 with boost at 4.0. I really have a hard time believing it wont sustain atleast 4ghz on all cores with decent cooling.

Some are pretty awesome such as Predator 360 can be purchased as an AIO, and it is rather nice albeit not cheap.
240 would be the most for most scenario as 360 is designed for CPU+GPU.
Guru3d, actual voltage was 1.35V rather than 1.3V.
http://www.guru3d.com/articles_pages/ek_predator_360_aio_cpu_gpu_liquid_cooling_review,1.html

Predator 240 AIO at Jayztwocents:


Cheers

 

[os2wiz]

What we do know:

Likely have Sandy or sub Sandy Instructions per Clock, as shown by PM single thread a 3.3/3.4Ghz 2500/2600K get about 1600 to Ryzen's 2050. There is a very causal link to that so must be Sandy clock vs clock level notwithstanding my 3.4ghz Haswell nets slightly lower scores.

It probably wont make it past 3.5ghz because of stuff.

Prime and Physics are now the benchmark of choice

All point to something like a Nahelem/Sandcheeks.

IPC can not becwritten in stone by one singlet threaded app. Ryzen 8 core 1700X or 1800X has been been benchmarked on several single threaded apps. Some it has been at Broadwell levels, some Haswell levels, other just below Haswelll levels, some at Ivy Bridge levels. On average it is a small notch above Haswell levels. It is time for you to stop spreading your mythology abot Ryzen IPC being Sandybridge levels that is nothing less than a crok of crap created by the infamous troll Juanrga. You can do better than that. I do not agree with those claiming Ryzen is at Star Lake levels,I think they are fanboys, but Ryzen has a competitive though lesser IPC to Intels best. The Kaby lake results are based on clock speed alone. Kaby Lake IPC is only about 1 to two percent above Star Lake. The second round of Ryzen chips later this year will have iterations in 4 core , 6 core, and 8 core that clock higher than the ones released now. The 4 and 6 core versions are made from cannabalized 8 cores and as a result are not as good clock speed wise as tge high end 1800X. In 3 to 4 months there willl be superior 4 and 6cores released on dedicated 4 and 6 core wafers. As the rocess is refined therevwill be fewer and fewer defects and under performing 8 cores and financially speaking that would be the best time to have dedicated 4 and 6 core wafers that will clock considerably higher than current skus.

 

[OrangeKhrush]

IPC can not becwritten in stone by one singlet threaded app. Ryzen 8 core 1700X or 1800X has been been benchmarked on several single threaded apps. Some it has been at Broadwell levels, some Haswell levels, other just below Haswelll levels, some at Ivy Bridge levels. On average it is a small notch above Haswell levels. It is time for you to stop spreading your mythology abot Ryzen IPC being Sandybridge levels that is nothing less than a crok of crap created by the infamous troll Juanrga. You can do better than that. I do not agree with those claiming Ryzen is at Star Lake levels,I think they are fanboys, but Ryzen has a competitive though lesser IPC to Intels best. The Kaby lake results are based on clock speed alone. Kaby Lake IPC is only about 1 to two percent above Star Lake. The second round of Ryzen chips later this year will have iterations in 4 core , 6 core, and 8 core that clock higher than the ones released now. The 4 and 6 core versions are made from cannabalized 8 cores and as a result are not as good clock speed wise as tge high end 1800X. In 3 to 4 months there willl be superior 4 and 6cores released on dedicated 4 and 6 core wafers. As the rocess is refined therevwill be fewer and fewer defects and under performing 8 cores and financially speaking that would be the best time to have dedicated 4 and 6 core wafers that will clock considerably higher than current skus.

Lol, my post was sarcasm, there are a few doom mongers here and there that I am tired of wasting my data on.

I know Juan well, after 3 years of his non stop garbage on Toms, I lost my chizz and laid out on him, got a perma ban for hostile and offensive abuse. It was worth it, all those badges and trophies were so worth it.

 

[CSI_PC]

I kind of agree,you're just throwing terms out there to scare him away. None of that will tell you how well this thing will overclock over its base clock. Obviously they can't hit 220w at 14nm, but Intel chips have shown they can go up to ~130w on a 91w spec quad at 14nm, and we know the core sizes are similar.

Fwiw, I have a degree in microelectronics and wrote a paper on maximum clock speed at a given node for my semiconductor physics 2 class and none of those terms were in my paper. Maybe that's why I only got an 80 on it? The electromagnetic field on the gate insulator was the limiting factor in my paper as exceeding a certain amount would damage it. Electromigration is a longer term reliability problem,not an out of the box clock limiter.

That was theoretical though,you'd have to know the layout of the chip and test thermals throughout the chip layout to see where localized hot spots are that can limit performance,which all of the research papers in the world won't help you with. It's why you get higher clocks with hyper threading turned off on an Intel vs on,even though everything else is exactly the same. Hyper threading isn't causing more electromigration or even higher gate leakage, it's causing more localized heat somewhere on the chip causing slightly lower clocks. You can read all the papers you want and it won't tell you how or if that will affect clocks without testing it as it's architecture related, not node related. Given the ryzen chips are a completely new architecture that nobody has flogged yet,there's no real way to determine it. Maybe they've figured out how to enable hyper threading with no overclocked penalty? Maybe it's worse? Who knows?

From a silicon-node-fab EE perspective to me the problem could come down to the design and hotspot modelling and layout-resolution, transistor density (I appreciate dGPU is different but AMD always went with a much higher transistor density than Nvidia), leakage (also includes Caches), and importantly the efficiency-scheduling of the Power Budget - Static/Dynamic Power consumption.
In many ways this puts the advantage with Intel as they own the fabs and these are integral to any CPU design, so I doubt we will see AMD competing equally on highest clocks in the context of OCing outside of Boost.
All that said, considering the price-performance tier they are looking very interesting and cautiously optimistic especially the 8 core models (needs to be seen just how much the issues with cutting cores has on the lower core models).
Cheers

 

[CentroX]

So we can buy this thing in 12 days yet we dont know anything really. Amd has all ryzen info as tight as area51.

Is it because of intel or does it suck?