Some bad news

[razor1]

Yet paper does not work well in the cpu socket and Ryzen is still selling quite well on Amazon. Course every time you say AMD is screwed the stock surges, so I am good with that.

Really in the last 6 months how well has AMD stock been doing, pretty flat man. Ever since Zen was released, AMD stock is doing shit. Yeah great for them, Boo making no money.

What kind of comment is that? shit its like you don't even buy or sell stocks and make these comments?

Anyone that liked this post, sorry for you, cause I would stay away for investing in stocks if this is your mentality.

 

[razor1]

Can you explain this please how is the GCN architecture tied to power consumption where is the proof that each GCN design only goal is tied to using more power. From what place do you pull your inside information please share?

It is obvious that you can not be taken seriously, in your signature you claim to have a RX 480 which is a low power version of the R9 290X.

That is pretty much all it is, GCN has evolved so little..... Yeah the r4x 480 has improved in some front end though put issues, but we won't see that unless they start becoming bottlenecks.

Perf/watt has been getting worse in ever generation since the first GCN architecture vs the competition, this is really easy to see because even at 300 watts Vega can reach the gtx 1080ti. At least with 275 watts on Fiji it got to the 980ti.

 

[razor1]

Razor, although market share may stay flat, the volumes have increased from what I saw.
Basically AMD can make more $ without gaining marketshare, if volumes increase to both... e.g. more sales and % doesn't show it.

There is a difference between a 8700k and Zen. One you can actually buy, in stock and at MSRP.
One you cannot. Let alone that 3600THz golden virgin ram lol.

How did market share stay flat when Intel is making more money after Zen came out for 2 quarters, Intel has been beating the market and their own estimates and previous year quarters and previous quarters! Volume in sales have a direct impact on marketshare don't you agree?

I don't care how any one tries to spin this, if Zen was so great and everyone is buying them, we would see Intel take a hit if not the first quarter, the second quarter and still we are not seeing that, we are seeing the opposite!

You got Gideon talking about stock prices and he doesn't know AMD's stock hasn't shifted after Zen's launch, just been bouncing up and down and pretty flat..... worse yet in that 6 months after Zen released, Intel stock doubled! Does it mean anything? No it means shit! Ridiculous comments with no backing or understanding of why things are happening in the stock market! Stock price has little to do with the health of company, its all speculation!

Come on guys, if you want to talk about AMD's finances at least look at Intel's books too, cause anything that will help AMD will hurt Intel! That is why investing in any of their stocks is damn easy. (as long as you understand the general market trends).

Wooped dee! I like AMD so I buy their stock and it doubled in the past year makes me and AMD look great!, Show me the rest of you damn portfolio, do you do that in ever stock you buy? Or is that just fuckin luck! Its bullish market in the past year in tech wow you can't a profit right now, you suck at trading stocks, show me in the past 3 years you portfolios and tell me you can do just the same thing in a down market and then we can talk.

 

[Pieter3dnow]

That is pretty much all it is, GCN has evolved so little..... Yeah the r4x 480 has improved in some front end though put issues, but we won't see that unless they start becoming bottlenecks.

Perf/watt has been getting worse in ever generation since the first GCN architecture vs the competition, this is really easy to see because even at 300 watts Vega can reach the gtx 1080ti. At least with 275 watts on Fiji it got to the 980ti.

If the design was called wookie or banana it would not have made a difference. That is the best engineers at AMD can do regardless of the name if the name would change suddenly everything would be al right in the universe, please ....

 

[N4CR]

How did market share stay flat when Intel is making more money after Zen came out for 2 quarters, Intel has been beating the market and their own estimates and previous year quarters and previous quarters! Volume in sales have a direct impact on marketshare don't you agree?

I don't care how any one tries to spin this, if Zen was so great and everyone is buying them, we would see Intel take a hit if not the first quarter, the second quarter and still we are not seeing that, we are seeing the opposite!

You got Gideon talking about stock prices and he doesn't know AMD's stock hasn't shifted after Zen's launch, just been bouncing up and down and pretty flat..... worse yet in that 6 months after Zen released, Intel stock doubled! Does it mean anything? No it means shit! Ridiculous comments with no backing or understanding of why things are happening in the stock market! Stock price has little to do with the health of company, its all speculation!

Come on guys, if you want to talk about AMD's finances at least look at Intel's books too, cause anything that will help AMD will hurt Intel! That is why investing in any of their stocks is damn easy.

Marketshare can stay flat while growth in sector happens e.g. sales volumes increase for both sides proportionally. You and I both know Zen has been selling well compared to last efforts. They both have strong products in each sector. What I'm also getting at is shares don't always directly reflect sales, as you have covered, especially if they doubled and there is nothing to show for it, especially in one part of Intel most here are talking about. They can also reflect other goings on, you do realize Intel has been diversifying into other areas outside of PC/fab related stuff? I can't comment on what I know personally about this due to restrictions but there are news items about this in other areas and lets just say the CPU/processor/fab market isn't their only game, this is their (Intel) long-term survival strategy. That previously core market has flattened out mostly and is why they are diversifying.

 

[razor1]

Marketshare can stay flat while growth in sector happens e.g. sales volumes increase for both sides proportionally. You and I both know Zen has been selling well compared to last efforts. They both have strong products in each sector. What I'm also getting at is shares don't always directly reflect sales, as you have covered, especially if they doubled and there is nothing to show for it, especially in one part of Intel most here are talking about. They can also reflect other goings on, you do realize Intel has been diversifying into other areas outside of PC/fab related stuff? I can't comment on what I know personally about this due to restrictions but there are news items about this in other areas and lets just say the CPU/processor/fab market isn't their only game, this is their (Intel) long-term survival strategy. That previously core market has flattened out mostly and is why they are diversifying.

That means Intel also gained, then there is no marketshare gain, there is a problem though, Intel gained more than AMD.....

Do you release Intel bottom line went up 10% while AMD's bottom line did 5%? That to me doesn't sound like AMD Zen is helping much outside of margins. If we are talking about stronger quarters that are helping both AMD and Intel out. Then AMD is still screwed, cause when weaker quarters hit it doesn't matter. AMD's zen would not have pushed Intel's sales we know that, Kaby Lake wasn't much different that Skylake, its not going to improve sales, so the only thing that would help sales right now for Intel is seasonal sales. Q1 and Q2 are usually weak, yet they beat the street. Worse yet the traditional drop that they have didn't happen as much. Intel exploded in profits just before and after Zen came out!

https://ycharts.com/companies/INTC/revenues

Do you see whats happening to Intel's revenue

Right now their year low is higher then last years high, and this is when Zen is here!

This is also after expenses so the past couple quarters they bought a few billion dollars worth of corporations.... not sure if it comes into this line item but it should.

 

[razor1]

If the design was called wookie or banana it would not have made a difference. That is the best engineers at AMD can do regardless of the name if the name would change suddenly everything would be al right in the universe, please ....

Its a over used or old design. Its not a "bad" design. Its just been used for too long.

 

[noko]

Razor1

• Trump effect

Sales went up for Do It Yourselfers mostly - initial impact, Dell, HP and Lenova OEM sells we have to see how that falls over time. Probably too early for those. I don't think we will know clearly the impact of Ryzen until next year but the signs look rather good to me. Next Financial Report should give better indicators. Ryzen is still at a very early stage for the Product Life Cycle, almost in the Introductory stage (For the Pro version yes, Mobile is not even at the Introductory stage except for some desktop versions brought to it) but I say it is now in the Growth stage for DIY. We are still looking at future releases with Ryzen. How well it matures and is maintained via effective updates is anyone's guess.

 

[razor1]

Razor1

• Trump effect

Sales went up for Do It Yourselfers mostly - initial impact, Dell, HP and Lenova OEM sells we have to see how that falls over time. Probably too early for those. I don't think we will know clearly the impact of Ryzen until next year but the signs look rather good to me. Next Financial Report should give better indicators. Ryzen is still at a very early stage for the Product Life Cycle, almost in the Introductory stage (For the Pro version yes, Mobile is not even at the Introductory stage except for some desktop versions brought to it) but I say it is now in the Growth stage for DIY. We are still looking at future releases with Ryzen. How well it matures and is maintained via effective updates is anyone's guess.

Yeah the bullish market is somewhat do to Trump but not entirely, companies like Intel, Google, Facebook, Apple, and many others don't like Trumps policies with Immigration, it affects H1B visa's and that is where most of the tech force is. So its actually a counter effect we are seeing.

DIY is where Ryzen is now but since the beginning of Q2, OEM's have picked up Ryzen so those are part of it too.

Added to this Intel just went through a major restructuring another factor that didn't even have a blimp in revenues. Where did all the money for severance packages go? Its in these last 2 or 3 Q's but because Intel has been beating everything they have stated they don't show up.

 

[noko]

Yeah the bullish market is somewhat do to Trump but not entirely, companies like Intel, Google, Facebook, Apple, and many others don't like Trumps policies with Immigration, it affects H1B visa's and that is where most of the tech force is. So its actually a counter effect we are seeing.

DIY is where Ryzen is now but since the beginning of Q2, OEM's have picked up Ryzen so those are part of it too.

Added to this Intel just went through a major restructuring another factor that didn't even have a blimp in revenues. Where did all the money for severance packages go? Its in these last 2 or 3 Q's but because Intel has been beating everything they have stated they don't show up.

Those buying stocks may careless about what the company thinks - Taxes or tax breaks, the burden of Obama Care era almost overt may look good to investors into these companies even if the company can't stand in general Trump (I doubt that though myself). If AMD is clearly in the black this round I would expect the stock to go up. Then we could have a major conflict in the Eastern region which can cause utter chaos (Intel would probably fare better in such a case, safer bet). Very complex puzzle with changing pieces all the time.

 

[razor1]

Those buying stocks may careless about what the company thinks - Taxes or tax breaks, the burden of Obama Care era almost overt may look good to investors into these companies even if the company can't stand in general Trump (I doubt that though myself). If AMD is clearly in the black this round I would expect the stock to go up. Then we could have a major conflict in the Eastern region which can cause utter chaos (Intel would probably fare better in such a case, safer bet). Very complex puzzle with changing pieces all the time.

Its most likely going to go into the black but just not much. I just don't see it swinging very fast.

 

[Pieter3dnow]

Its a over used or old design. Its not a "bad" design. Its just been used for too long.

So can you explain to me why AMD can do certain optimizations to it, where you get lower power from R9 290X to RX 480, on other stuff they still struggle it is has very little to do GCN as a whole ...

 

[razor1]

So can you explain to me why AMD can do certain optimizations to it, where you get lower power from R9 290X to RX 480, on other stuff they still struggle it is has very little to do GCN as a whole ...

What the node doesn't come into play? Does that have anything to do with design, not really.

Was the G80 a bad design? But by the time the 2xx line rolled out it looked old didn't it? New nodes only help so much when architecture is being bottlenecked else where, the node can't help anymore. Yeah the 2xx was fast chip but compared to the competition it was a bigger chip, much bigger and barely kept up with the HD4xxx line and used more power, much more.

This is the same thing that has happened to GCN. Fiji same size chip and barely kept up 980ti, Vega much bigger chip and can't keep up with GP102, and barely keeps up with a much smaller chip GP104 and it uses much more power than either of these two chips.

AMD dropped two node sizes to get a 60% improvement in perf/watt, nV did that within the same node from Keplar to Maxwell, that is what really helps in perf/watt, the architecture. The node will give power savings but it can't stop a chip from being bottlenecked because of design choices.

 

[Araxie]

So can you explain to me why AMD can do certain optimizations to it, where you get lower power from R9 290X to RX 480, on other stuff they still struggle it is has very little to do GCN as a whole ...

The rx 480 its more similar in performance to the r9 290 than the 290x, the lower power consumption comes from the first fact: smaller node and billions of transistor spent only to increase clock.. even moderate changes on GCN didn't offered major changes in real world gaming beyond couple of games where geometry output was bottlenecking before as fallout 4, r9 280x to r9 380x was a major disappoint, the 280x was still better 99% of the tkme due wider bus and bandwidth, so all those optimizations from first gen GCN to third gen GCN where did go? And from 1.0 to 1.2 where presented the biggest changes to GCN gen to gen.. GCN was still the same power hungry, bandwidth starved architecture 2 revisions later, overclocked was even worse, 380X pulled more power than overclocked 280x and performed worse.. only reason RX 480 perform close to the r9 290 it's because due clocks. Same happened from Fury to Vega.. all *optimizations* are higher clocks.. power consumption it's also worse.. after a certain point there aren't too many changes you can apply to an architecture to keep it relevant.. most performance gains on GCN always comes due to clocks not due architectural changes. So what exactly optimization are you referring to?

 

[Gideon]

Really in the last 6 months how well has AMD stock been doing, pretty flat man. Ever since Zen was released, AMD stock is doing shit. Yeah great for them, Boo making no money.

What kind of comment is that? shit its like you don't even buy or sell stocks and make these comments?

Anyone that liked this post, sorry for you, cause I would stay away for investing in stocks if this is your mentality.

Oh look your hurt and felt the need to lash out. You keep forgetting before you got banned from the AMD forum that you said Ryzen was garbage, would not compete and last I checked AMD stock surged after it was released. It has sustained it's massive gains and that is a good thing for any stock, if you actually knew anything. I buy and sell tech stock all the time and I keep others just for the dividends, I just wish I had bought more Google stock when it was first offered. I mean hell Tesla has never turned a freaking profit and you could have made a fortune off their stock value, it's you that has no clue how the stock market works. Profit and company stock value really have little to do with each other, speculation and buzz have more to do with it. I am sure you will have some cranky retort for me, but literally I could not care, your doing the same thing in the AMD Video Card forum you used to do in the AMD cpu forum and that is always berate AMD.

 

[razor1]

Oh look your hurt and felt the need to lash out. You keep forgetting before you got banned from the AMD forum that you said Ryzen was garbage, would not compete and last I checked AMD stock surged after it was released. It has sustained it's massive gains and that is a good thing for any stock, if you actually knew anything. I buy and sell tech stock all the time and I keep others just for the dividends, I just wish I had bought more Google stock when it was first offered. I mean hell Tesla has never turned a freaking profit and you could have made a fortune off their stock value, it's you that has no clue how the stock market works. Profit and company stock value really have little to do with each other, speculation and buzz have more to do with it. I am sure you will have some cranky retort for me, but literally I could not care, your doing the same thing in the AMD Video Card forum you used to do in the AMD cpu forum and that is always berate AMD.

LOL I got banned cause people like you didn't like it when I stated Ryzen isn't all that. Its decent, but nothing that will give you wet dreams over, maybe it does for you What ever floats your boat right?

There was massive a surge in stock price, lets see.

Here ya go.

Year to date data for AMD stock prices

https://finance.yahoo.com/chart/AMD

When was Ryzen released Feb


Why did AMD's stock price radically drop after Feb, all those bullish investors just wanted their money! Right, no its large institutions that had options cashing out!

Why hasn't AMD's stock price gone up after Ryzen's release, if they are making a butt load of sales, it should reflect correct? Why is still just going up and down at the same levels? Cause investors are looking at it and need more then what AMD is showing them on their bottom line. This is what I stated, Ryzan ain't going to do shit for AMD cause its not giving anything more than Intel and it has a butt load of problems with CCX cross talk!

You going to sit here and deny that? Why does Epyc get a royal beat down in multithreaded tasks that require localization? You going to sit here and tell me server companies don't care about things like that? Since you are a car fan, that is like taking a V12 to the race but can only use 4 cylinders. You think companies are going to pay thousands of dollars for a chip that works 25% of the time at 100% performance? Great "AMD's Back" in the DIY market where the only people that purchase such systems are cash bound! YEAH! makes AMD look really good for the cash strapped!

Come again, want to try that analysis again?

I wasn't talking about before Ryzen was released, I was talking about after, there has been no change to its stock price, because now AMD must deliver on all the BS they have been spouting before, if they don't, you will see their stock price drop, because at the moment there is no large institutional investors (well large enough to make the market sway) that are bullish with AMD, AMD needs to show they have done what they have hyped. Its been two Q's and still nothing, not a good sign man (stagnation is never a good sign, that means there is no movement in either direction, it makes the stock not worth holding on to in the long run, cause people can make money on other stocks that have enough volatility either way, lets see what this month brings, if Intel's predictions are correct from their end then AMD hasn't hit their hype, and this is why AMD has been silent on forecasting their results for he past two Q's, its too murky to know for them, its not up to them, its up to the OEM's and other companies to make the rest of AMD's promises reality and they are also in business with Intel, so ultimately, its also Intel.

Before you sit here and tell me anything more about stocks and their prices, read this

https://www.thebalance.com/the-investor-vs-speculator-358139

 

[Hakaba]

I have a few AMD die hard fans at work that dropped a few thousand into AMD stock when it was 15 ish dollars, not when it was sub 10, when it was 15,,,, Citing it was about to go to the moon so they could buy their lambo, hell Ryzen was about to release so was Vega! A few banks/other sources were also estimating a 40%+ increase in stock value before or shortly after product launch....

Yeah... I am not the best when it comes to stocks, I’ve seen crazy increases on companies that have little to no value only because people thought it was worth something. Right now I don’t foresee AMD shooting up, not till they enter other markets, or see worth outside the eyes of their fanbase.

Not saying AMD is worthless, but Intel and NVIDIA are the known, sold, brands, to the majority.

 

[Gideon]

LOL I got banned cause people like you didn't like it when I stated Ryzen isn't all that. Its decent, but nothing that will give you wet dreams over, maybe it does for you What ever floats your boat right?

There was massive a surge in stock price, lets see.

Here ya go.

Year to date data for AMD stock prices

https://finance.yahoo.com/chart/AMD

When was Ryzen released Feb


Why did AMD's stock price radically drop after Feb, all those bullish investors just wanted their money! Right, no its large institutions that had options cashing out!

Why hasn't AMD's stock price gone up after Ryzen's release, if they are making a butt load of sales, it should reflect correct? Why is still just going up and down at the same levels? Cause investors are looking at it and need more then what AMD is showing them on their bottom line. This is what I stated, Ryzan ain't going to do shit for AMD cause its not giving anything more than Intel and it has a butt load of problems with CCX cross talk!

You going to sit here and deny that? Why does Epyc get a royal beat down in multithreaded tasks that require localization? You going to sit here and tell me server companies don't care about things like that? Since you are a car fan, that is like taking a V12 to the race but can only use 4 cylinders. You think companies are going to pay thousands of dollars for a chip that works 25% of the time? Great "AMD's Back" in the DIY market where the only people that purchase such systems are cash bound! YEAH! makes AMD look really good for the cash strapped!

Come again, what to try that analysis?

Sure let me sum that up for ya... Razor1 Hates AMD.

Also in a one year time span it's gone from $6.75 to $15.20 in February of this year to a current $14.22, Yeah that current looks closer to the peak then the bottom, the sad part is you actually believe your own Hyperbole. I owned AMD stock when it was $2 a share and sold it around 12 bucks a share. Epyc just hit the market will take 6 months to see much on it and if they get Server wins they will see a nice uptick in stock value. This however is not the AMD cpu forum nor the Stock Market forum. You and a few others tho need to form the CCX haters club, perhaps a picture of Superglue as a avatar.

 

[razor1]

Sure let me sum that up for ya... Razor1 Hates AMD.

Also in a one year time span it's gone from $6.75 to $15.20 in February of this year to a current $14.22, Yeah that current looks closer to the peak then the bottom, the sad part is you actually believe your own Hyperbole. I owned AMD stock when it was $2 a share and sold it around 12 bucks a share. Epyc just hit the market will take 6 months to see much on it and if they get Server wins they will see a nice uptick in stock value. This however is not the AMD cpu forum nor the Stock Market forum. You and a few others tho need to form the CCX haters club, perhaps a picture of Superglue as a avatar.

Yep a guy that spent more on AMD products then you have in a life time hates them, I hate their products now cause they just can't do what nV's can do, sorry that is reality, everything from mining to gaming to 3d, anything AMD can do, nV cards do better!

Oh one year span you want to look at nV at the same time?

They grew more than AMD too.

Its a fuckin bull market what do you expect! What you think going from 2ish to 14 is great, how about 20 to 200? How about that? Didn't matter who you put you money into you get a huge ass return, you would have made more money if you invested in nV!

Hyper bole on Eypc, those weren't hyper bole figures man, yeah Eypc looses more than 50% of its performance when doing data locality. Its huge, and the programs that live off of that, are pretty much anything to do with the web, large databases. That is 75% of the server market if not more, I was being generous when I stated 25% of the time. Its more like 90% of the time Eypc will fall flat on its face! Talked to a our server programmers, Eypc can only get its performance back when recording an application with special function: NUMA interleave all, which pretty much cuts effectively based on the program down to one CXX! So if you have a program or database that requires more than 8 database nodes you are screwed.

You don't know any of this. Head in sand is a great approach to investing! This is why AMD hasn't been forecasting anything, they know its going to damn hard to do anything in servers, now in desktops they aren't getting anywhere. Intel is expecting nothing from AMD as well cause they figure their numbers are fairly stable.

AMD as company has always been optimistic about their fiscal quarters, and usually don't make them, and now they are uncertain, and you have faith in them when they don't have faith in themselves?

Good for you, they need a new PR guy, they should hire you!

oh when you interview for AMD, don't tell them that the tech sector grew 60% this year ok?

 

[Pieter3dnow]

The rx 480 its more similar in performance to the r9 290 than the 290x, the lower power consumption comes from the first fact: smaller node and billions of transistor spent only to increase clock.. even moderate changes on GCN didn't offered major changes in real world gaming beyond couple of games where geometry output was bottlenecking before as fallout 4, r9 280x to r9 380x was a major disappoint, the 280x was still better 99% of the tkme due wider bus and bandwidth, so all those optimizations from first gen GCN to third gen GCN where did go? And from 1.0 to 1.2 where presented the biggest changes to GCN gen to gen.. GCN was still the same power hungry, bandwidth starved architecture 2 revisions later, overclocked was even worse, 380X pulled more power than overclocked 280x and performed worse.. only reason RX 480 perform close to the r9 290 it's because due clocks. Same happened from Fury to Vega.. all *optimizations* are higher clocks.. power consumption it's also worse.. after a certain point there aren't too many changes you can apply to an architecture to keep it relevant.. most performance gains on GCN always comes due to clocks not due architectural changes. So what exactly optimization are you referring to?

Well I ran the FF XV benchmark and got about the same scores as Cageymaru on his setup (R9 290X vs RX 480)
You are trying to tell me something but I see no prove that suggests that GCN is to blame for the power/performance issues.
I gave a simple example and you are claiming it is due to smaller node and in reality if that was the case the difference was approx. 8 nm (from 28nm to some finfet process from GF called 14nm but in reality closer to 20nm). That would make little sense since there no other results that other products show in the same way from going to 28nm to GF "14nm finfet" .

All you claim is that the design is called GCN there for it has these issues.

 

[razor1]

Well I ran the FF XV benchmark and got about the same scores as Cageymaru on his setup (R9 290X vs RX 480)
You are trying to tell me something but I see no prove that suggests that GCN is to blame for the power/performance issues.
I gave a simple example and you are claiming it is due to smaller node and in reality if that was the case the difference was approx. 8 nm (from 28nm to some finfet process from GF called 14nm but in reality closer to 20nm). That would make little sense since there no other results that other products show in the same way from going to 28nm to GF "14nm finfet" .

All you claim is that the design is called GCN there for it has these issues.

GCN isn't to blame for the perf/watt? Seriously? What do you blame it on then, when a rx 480 which uses as much juice as a 1070 gets 30% less performance? Similar nodes......

come again?

The rx 480 in this chart has the same perf/watt ratio as a gtx 980 ti one is a much bigger chip and is on an older node........

28 nm node needs more voltage man. Much more, anyone want to pull up data on voltage vs. node size?

https://www.extremetech.com/computi...samsung-globalfoundries-intel-gear-7nm-beyond

please read.

Even better

https://www.anandtech.com/show/11337/samsung-and-tsmc-roadmaps-12-nm-8-nm-and-6-nm-added

So AMD going from 28nm for GCN (hawaii) to 14nm for GCN (polaris), almost the same transistor counts, so fairly viable comparison. Should have gotten 60% in power savings or getting 40% increase in performance.

Did we get that?

You yourself stated that didn't really happen, they got a 50% in power savings, but performance wise, its a wash, same performance.

 

[Pieter3dnow]

GCN isn't to blame for the perf/watt? Seriously? What do you blame it on then, when a rx 480 which uses as much juice as a 1070 gets 30% less performance? Similar nodes......



come again?

The rx 480 in this chart has the same perf/watt ratio as a gtx 980 ti one is a much bigger chip and is on an older node........

They were always behind in the scope of power/performance . You can show me benchmarks until you are blue in the face but it does not change that you can blame it on GCN. The design of the chip is power hungry it always has been it does not in any way shape or form prove that the architecture it self is to blame for that.
Keep working on your post count buddy ...

 

[razor1]

They were always behind in the scope of power/performance . You can show me benchmarks until you are blue in the face but it does not change that you can blame it on GCN. The design of the chip is power hungry it always has been it does not in any way shape or form prove that the architecture it self is to blame for that.
Keep working on your post count buddy ...

So you think what AMD and GF stated 60% decrease in power usage or 40% increase in performance, which they did get with GCN, was what? Its exactly what they got same performance with a 60% drop in power draw. actually it was more like 50% in power draw so they missed the node estimations. Ok 10% no big deal.

You too work on your post counts, doesn't help when you didn't read everything man.

Now lets talk about Vega, Vega should have gotten that right?

But wait it didn't, it actually went up in power usage and got the 40% increase in performance. Ouch. how do you rationalize that if it wasn't for an architectural fault? Was the node bad, too much leakage, so they needed overvolting to keep signal cohesion? No wait, voltage they are using in Vega is within limits of the node. Hmm ok, what else you want to try?

 

[Ramar]

Guys my uncle at Nintendo says the Wii U sucked so Nintendo is going software only.

 

[Chimpee]

The design of the chip is power hungry it always has been it does not in any way shape or form prove that the architecture it self is to blame for that.

I feel like this sentence contradict itself.

 

[Araxie]

Well I ran the FF XV benchmark and got about the same scores as Cageymaru on his setup (R9 290X vs RX 480)
You are trying to tell me something but I see no prove that suggests that GCN is to blame for the power/performance issues.
I gave a simple example and you are claiming it is due to smaller node and in reality if that was the case the difference was approx. 8 nm (from 28nm to some finfet process from GF called 14nm but in reality closer to 20nm). That would make little sense since there no other results that other products show in the same way from going to 28nm to GF "14nm finfet" .

All you claim is that the design is called GCN there for it has these issues.

you have to consider that cageymaru overclock always his shits as much as possible. and for sure a heavy overclocked RX 480 on water can match a stock 290X and for sure with a high overclock an RX 480 will also be close to the same power consumption..

The design of the chip is power hungry it always has been it does not in any way shape or form prove that the architecture it self is to blame for that.

What the hell man, explain me how the architecture of a chip is not the same as design of the chip.. man, GCN it's power hungry it's the same as say *the design of the chip is power hungry it always has been*.. Im really confused on what are you trying to say here because it's completely contradictory.

 

[tungt88]

Guys my uncle at Nintendo says the Wii U sucked so Nintendo is going software only.

Haha, well admittedly the demise of the Sega Dreamcast almost 2 decades ago was the "nail in the coffin" that led to Sega becoming software-only. Mediocre sales in Japan, but a then-incredible 500k+ Dreamcast consoles sold in a two-week period in the US. Still couldn't make up for five consecutive years of losses, man. That, combined with the bad rep Sega got earlier with Saturn/32X/Sega CD, and the upcoming PS2 really hurt Sega badly (ate up a lot of the goodwill from the very successful Genesis release). Shame, too, because the Dreamcast was actually a really awesome and well-thought-out console (my cousins had 'em).

 

[noko]

Folks, perf/watt also depends upon what type of workload the GPU is doing. In other words Vega with most current games type workloads fair worse than Nvidia for perf/w. That does not mean Vega cannot beat Nvidia in other types of workloads, like compute type workloads for perf/w. Best to match up the processor strengths to the best workloads to optimize. Just hounding GCN or making a generalization on perf/watt based on a limited scenarios can fall short in the overall view. Now if one is totally concern only with current games - Nvidia has a clear advantage. Future games may be different depending if the strengths of GCN can be used effectively or not.

 

[Hakaba]

Oh one year span you want to look at nV at the same time?

What you think going from 2ish to 14 is great, how about 20 to 200? How about that? Didn't matter who you put you money into you get a huge ass return, you would have made more money if you invested in nV!

Tried explaining that to someone before, who turned around and something along the line, that AMD is still better because growth percentage was higher than Nvidias...

Either would have netted returns, one offered better returns and dividends.

 

[razor1]

Folks, perf/watt also depends upon what type of workload the GPU is doing. In other words Vega with most current games type workloads fair worse than Nvidia for perf/w. That does not mean Vega cannot beat Nvidia in other types of workloads, like compute type workloads for perf/w. Best to match up the processor strengths to the best workloads to optimize. Just hounding GCN or making a generalization on perf/watt based on a limited scenarios can fall short in the overall view. Now if one is totally concern only with current games - Nvidia has a clear advantage. Future games may be different depending if the strengths of GCN can be used effectively or not.

it can't beat it in compute only work loads either, the only thing holding back Psacal gaming cards in eth only mining, is GDDR5x, we can see that with P100, with HBM 2 the same memory Vega uses, a whopping 69mhs, that is 30% higher then Vega, both cards at stock, granted can't get a P100 card for less than 6k but still, that shows the real performance of nV's compute capabilities. Gaming cards kick ass at mining all other alt coins, like at Zcash's hashrates on Pascal vs Vega. 1080ti, gets 650 sols or so non overclocked what is Vega doing? 475 sols..... big difference even with a shader disadvantage for Pascal. I can get 2.5-3 bucks per day with one 1080ti inclusive of power, Vega 1 to 2 bucks a day inclusive of power.

What is the difference the shader units are not being used effectively on Vega (this is not the software either cause older GCN's don't do well in Zcash comparatively to nV cards), the gtx 1080 gets the same sol rates as Vega 64......

It doesn't stop there, any alt coin per flop amount, nV card do better in that metric. These are pure compute. What can we say about Vega's perf/watt in compute only tasks? Kinda sucks. The reason why nV hardware wasn't in mining till the past year or so, wasn't because they weren't capable of it, its just the cuda mining software wasn't there to fully expose nV's cards full potential, we know nV doesn't care about OpenCL lol.

 

[razor1]

Tried explaining that to someone before, who turned around and something along the line, that AMD is still better because growth percentage was higher than Nvidias...

Either would have netted returns, one offered better returns and dividends.

Yeah people have tunnel vision, they don't understand the market is based on feelings, if one tech company starts doing good all others in that specific market also get pushed up. In this case it started with nV, I remember when it was @ 9 bucks like 3 years ago and told my dad to buy a ton of it and sit on it, he did, he also bought AMD too later on when it was around 2 or 3 bucks. Also bought Intel a year ago.

Its a fuckin no brainier when the market is bullish and stock prices are lower than what the actual company is worth. Its knowing when to stop that takes skill, being too greedy and not knowing the markets that your investing in will screw investors.

I remember an investor from NYC *Brooklyn* in the late 90's, he was featured on American Greed, smart fuckin guy and hit it lucky in a bull market so he opened up his own trading firm down in Florida, did great when the market was hot, not so well after the market tanked, then he started a ponzi scheme to get the money back that he lost.....

it was all because he didn't understand the forces of the market, not because he didn't know what he was investing in. he knew the companies he invested in, very well, he could tell you things about what the CEO was doing on weekly basis.

 

[noko]

it can't beat it in compute only work loads either, the only thing holding back Psacal gaming cards in eth only mining, is GDDR5x, we can see that with P100, with HBM 2 the same memory Vega uses, a whopping 69mhs, that is 30% higher then Vega, both cards at stock, granted can't get a P100 card for less than 6k but still, that shows the real performance of nV's compute capabilities. Gaming cards kick ass at mining all other alt coins, like at Zcash's hashrates on Pascal vs Vega. 1080ti, gets 650 sols or so non overclocked what is Vega doing? 475 sols..... big difference even with a shader disadvantage for Pascal. I can get 2.5-3 bucks per day with one 1080ti inclusive of power, Vega 1 to 2 bucks a day inclusive of power.

What is the difference the shader units are not being used effectively on Vega (this is not the software either cause older GCN's don't do well in Zcash comparatively to nV cards), the gtx 1080 gets the same sol rates as Vega 64......

It doesn't stop there, any alt coin per flop amount, nV card do better in that metric. These are pure compute. What can we say about Vega's perf/watt in compute only tasks? Kinda sucks. The reason why nV hardware was in mining till the past year or so, wasn't because they weren't capable of it, its just the cude mining software wasn't there to fully expose nV's cards full potential, we know nV doesn't care about OpenCL lol.

Maybe I will put that to the test, using some compute stuff between 1080Ti and Vega 64 - see who wins out on the test and then perf/w. Make it so anyone can do a peer review of the results. Vega has the horsepower so to speak, see if it can be transferred to the pavement.

 

[Gideon]

Yep a guy that spent more on AMD products then you have in a life time hates them, I hate their products now cause they just can't do what nV's can do, sorry that is reality, everything from mining to gaming to 3d, anything AMD can do, nV cards do better!

Oh one year span you want to look at nV at the same time?

They grew more than AMD too.

Its a fuckin bull market what do you expect! What you think going from 2ish to 14 is great, how about 20 to 200? How about that? Didn't matter who you put you money into you get a huge ass return, you would have made more money if you invested in nV!

Hyper bole on Eypc, those weren't hyper bole figures man, yeah Eypc looses more than 50% of its performance when doing data locality. Its huge, and the programs that live off of that, are pretty much anything to do with the web, large databases. That is 75% of the server market if not more, I was being generous when I stated 25% of the time. Its more like 90% of the time Eypc will fall flat on its face! Talked to a our server programmers, Eypc can only get its performance back when recording an application with special function: NUMA interleave all, which pretty much cuts effectively based on the program down to one CXX! So if you have a program or database that requires more than 8 database nodes you are screwed.

You don't know any of this. Head in sand is a great approach to investing! This is why AMD hasn't been forecasting anything, they know its going to damn hard to do anything in servers, now in desktops they aren't getting anywhere. Intel is expecting nothing from AMD as well cause they figure their numbers are fairly stable.

AMD as company has always been optimistic about their fiscal quarters, and usually don't make them, and now they are uncertain, and you have faith in them when they don't have faith in themselves?

Good for you, they need a new PR guy, they should hire you!

oh when you interview for AMD, don't tell them that the tech sector grew 60% this year ok?

It amazes me how much you know about me and what I have bought or not. Of course you have done everything and better then everyone else and own this huge house, nice cars and yet you sit here on a forum telling everyone how much better and right you are Yes Nvidia did better, Google also did awesome since its initial offering, only matters if you actually bought it and I have, but like I said this isn't the Stock Market forum. Somehow I think anyone looking at server hardware knows what they need and the software they run, I dont think they call you up for a consult on what they need so whining about a weak spot in AMD Epyc is kind of silly as it excels at many other tasks. Hey I am shocked your ego allowed you to say I would be good at something, tho PR was never my strong suit. Failure analysis is what I am good at and helping a engineer correct issues with their designs. You love to talk down and belittle anyone that disagrees with you, while I went out and had fun with my friends today (friends: are people that like and enjoy your company, cause I figure you didn't know).

 

[razor1]

It amazes me how much you know about me and what I have bought or not. Of course you have done everything and better then everyone else and own this huge house, nice cars and yet you sit here on a forum telling everyone how much better and right you are Yes Nvidia did better, Google also did awesome since its initial offering, only matters if you actually bought it and I have, but like I said this isn't the Stock Market forum. Somehow I think anyone looking at server hardware knows what they need and the software they run, I dont think they call you up for a consult on what they need so whining about a weak spot in AMD Epyc is kind of silly as it excels at many other tasks. Hey I am shocked your ego allowed you to say I would be good at something, tho PR was never my strong suit. Failure analysis is what I am good at and helping a engineer correct issues with their designs. You love to talk down and belittle anyone that disagrees with you, while I went out and had fun with my friends today (friends: are people that like and enjoy your company, cause I figure you didn't know).

Why are talking about intial offering of google now? What does IPO's have to do with what we are talking. What you can't stay one the topic because........

Epyc doesn't out do Intel's offerings in those other tasks, they are very close in those other tasks! Thats the point!

You get the same shit as Intel but Intel doesn't' have any fuckin issues with database localization!

Is that really that hard to understand?

Its Ryzen for Servers, same fuckin problems that Ryzen has just at a bigger scale!

Wow you just sit there and think AMD is so damn great when they still can't come back after what a decade?

You specifically stated after Ryzen's launch AMD is going to take it to Intel, what did I say, its going to take years before they get back to what they did with Athlon 64, it took a year or two before AMD got traction with A64 once they got traction then Intel started the illegal crap. Its been two quarters have we seen anything yet?

Things don't turn on a dime when there are contracts to uphold!

Only a person that sits at his computer and has nothing else better to do wouldn't understand that.

Failure analysis so you must know some actuary work, let me give you a problem

A company sells 0, 1, or 2 products on any day. When selling a product,the company tries to persuade the customer to buy an extended warranty for the product. Let X denote the number of products sold in a given day, and let Y denote the number of extended warranties sold.
P(X = 0, Y = 0) = 1 / 6
P(X = 1, Y = 0) = 1/12
P(X = 1, Y = 1) = 1 /6
P(X = 2, Y = 0) = 1 /12
P(X = 2, Y = 1) = 1 /3
P(X = 2, Y = 2) = 1/6

What is the variance of X..

Go for it man, you are failure analyst engineer, you should know this stuff. Your data is used for the expected outcome of the warranty sale and the warranty price.

Do you want some actuary questions about investing on the stock market? Yeah I've done math and finance courses that are way past graduate level. So you can sit here and try to convince me that AMD's stock price is solely based on their performance, but it isn't going to work because I KNOW better, I have been investing in stocks since I was 16 and know what makes a good trader and great trader over the guy that "thinks" they know something when working at investment banks. You need to know everything otherwise the chance of making a mistake even a small one can take a person making profit to making a loss.

Have you ever been on the receiving end of trader investment bank when they need something to get done ASAP in their databases or functions when their models aren't coming up right? I have never heard so many Fbombs in one sentence. You want to try things like what you say to these guys, what they would think of you? This is why all those BS analysts are just BS, they aren't the ones trading, not only that they have agendas for saying something!

Yet you have idiots that write up their things on AMD and nV and other companies on seeking alpha thinking they know something yet they don't.

Anyone getting stock advice from stuff like that without doing their own damn research are just going with the wave, if it goes up they go up, if it goes down they go down.

Actuary stock question a simple one

For a two-period binomial model, you are given:

Each period is one year.

The current price for a non dividend-paying stock is 20.

u= 1.2840, where u is one plus the rate of capital gain on the stock per period if the stock price goes up.

d =0.8607, where d is one plus the rate of capital loss on the stock per period if the stock price goes down.

The continuously compounded risk-free interest rate is 5%.

Calculate the price of an American call option on the stock with a strike price of 22.

This one is from a test I took in oh 2011 I think, it was to get certification for certain things I was doing at Credit Suisse.

If you think I know you didn't spend 60k on AMD products in your life time, yeah I know you didn't......

What reason would you need to do that ever, you aren't a miner, at least not as business. Why would anyone spend that much money, that's what in 10 years 6k on AMD products? Do you do that? How old are you 40 years old, ok 20 years of buying AMD products, still gets you 3k for AMD only products. Have you done that yet?

 

[_mockingbird]

Can you explain this please how is the GCN architecture tied to power consumption where is the proof that each GCN design only goal is tied to using more power. From what place do you pull your inside information please share?

It is obvious that you can not be taken seriously, in your signature you claim to have a RX 480 which is a low power version of the R9 290X.

That's what is called a straw man.

Regardless, it is safe to say that Radeon RX Vega 64 is the most power-hungry single-GPU ever released.

 

[Pieter3dnow]

That's what is called a straw man.

Regardless, it is safe to say that Radeon RX Vega 64 is the most power-hungry single-GPU ever released.

You just pulled a dirty tactic by posting a picture of a benchmark without the specification which driver setting was used because it might as well been overclocked and the performance instead of default setting or power saving . And you call me straw man .

Never can you quote a message where I say that the power usage is good either.

Why don't you link any post of me where I say these things. Instead you post a picture of something that only proves that RX Vega is not power efficient which no one denies.

For the record I'm not asking you to explain what I already know , I'm asking to explain how this is tied to GCN (I'm not asking you to show me a pack of butter I'm asking you to show me how the butter is made and explain the process...)

 

[Pieter3dnow]

I feel like this sentence contradict itself.

Why don't you take the time and stroll down the AMD youtube channel find any video that has been made about GCN. Then you might be a little wiser..
If GCN was so power hungry then why would the same GCN use less power on other cards that does make sense to you ?

 

[razor1]

Why are talking about intial offering of google now? What does IPO's have to do with what we are talking. What you can't stay one the topic because........

Epyc doesn't out do Intel's offerings in those other tasks, they are very close in those other tasks! Thats the point!

You get the same shit as Intel but Intel doesn't' have any fuckin issues with database localization!

Is that really that hard to understand?

Its Ryzen for Servers, same fuckin problems that Ryzen has just at a bigger scale!

Wow you just sit there and think AMD is so damn great when they still can't come back after what a decade?

You specifically stated after Ryzen's launch AMD is going to take it to Intel, what did I say, its going to take years before they get back to what they did with Athlon 64, it took a year or two before AMD got traction with A64 once they got traction then Intel started the illegal crap. Its been two quarters have we seen anything yet?

Things don't turn on a dime when there are contracts to uphold!

Only a person that sits at his computer and has nothing else better to do wouldn't understand that.

Failure analysis so you must know some actuary work, let me give you a problem

A company sells 0, 1, or 2 products on any day. When selling a product,the company tries to persuade the customer to buy an extended warranty for the product. Let X denote the number of products sold in a given day, and let Y denote the number of extended warranties sold.
P(X = 0, Y = 0) = 1 / 6
P(X = 1, Y = 0) = 1/12
P(X = 1, Y = 1) = 1 /6
P(X = 2, Y = 0) = 1 /12
P(X = 2, Y = 1) = 1 /3
P(X = 2, Y = 2) = 1/6

What is the variance of X..

Go for it man, you are failure analyst engineer, you should know this stuff. Your data is used for the expected outcome of the warranty sale and the warranty price.

Do you want some actuary questions about investing on the stock market? Yeah I've done math and finance courses that are way past graduate level. So you can sit here and try to convince me that AMD's stock price is solely based on their performance, but it isn't going to work because I KNOW better, I have been investing in stocks since I was 16 and know what makes a good trader and great trader over the guy that "thinks" they know something when working at investment banks. You need to know everything otherwise the chance of making a mistake even a small one can take a person making profit to making a loss.

Have you ever been on the receiving end of trader investment bank when they need something to get done ASAP in their databases or functions when their models aren't coming up right? I have never heard so many Fbombs in one sentence. You want to try things like what you say to these guys, what they would think of you? This is why all those BS analysts are just BS, they aren't the ones trading, not only that they have agendas for saying something!

Yet you have idiots that write up their things on AMD and nV and other companies on seeking alpha thinking they know something yet they don't.

Anyone getting stock advice from stuff like that without doing their own damn research are just going with the wave, if it goes up they go up, if it goes down they go down.

Actuary stock question a simple one

For a two-period binomial model, you are given:

Each period is one year.

The current price for a non dividend-paying stock is 20.

u= 1.2840, where u is one plus the rate of capital gain on the stock per period if the stock price goes up.

d =0.8607, where d is one plus the rate of capital loss on the stock per period if the stock price goes down.

The continuously compounded risk-free interest rate is 5%.

Calculate the price of an American call option on the stock with a strike price of 22.

This one is from a test I took in oh 2011 I think, it was to get certification for certain things I was doing at Credit Suisse.

If you think I know you didn't spend 60k on AMD products in your life time, yeah I know you didn't......

What reason would you need to do that ever, you aren't a miner, at least not as business. Why would anyone spend that much money, that's what in 10 years 6k on AMD products? Do you do that? How old are you 40 years old, ok 20 years of buying AMD products, still gets you 3k for AMD only products. Have you done that yet?

Guys anyone interested to know how that last question works, learn it I mean it, if you are a serious investor, all option data has to be disclosed, so you do have access to option data and when the calls come in. If you have that data and know how to to do this problem you will know how the big institutions work. You would have know AMD's stock price was going to drop right after Ryzen's launch, cause it was already in the books. Gideon, this is why I asked you why did it drop, you don't know these things. If you did, you would have shorted or did a put on AMD and would get with a 10% accuracy of hitting, instead of what doubling or tripling your money you could have gotten quadruple.

See information is freely available Gideon

http://www.nasdaq.com/symbol/amd/option-chain

Although I can solve the problem above, I don't play with options, cause well that 10% risk is more then I'm willing to do. I like it when my risk is less than 5%, but for investors like you, who don't know these things 10% risk is nothing for you, without knowing these things your risk is probably well above 25%.

 

[razor1]

Why don't you take the time and stroll down the AMD youtube channel find any video that has been made about GCN. Then you might be a little wiser..
If GCN was so power hungry then why would the same GCN use less power on other cards that does make sense to you ?

Don't try to rewrite history, when AMD needs 300 watts to do the same thing as a 180 watt card yeah its not just power hungry.

What are you talking about, cut down version of GCN that have less clocks? Of course they will use less power, but then again to attain gtx 1070 performance levels they are still using 210+ watts vs a 1070 which is what? 150 watts?

Oh on this note what ever happened to Vega Nano, we saw them give it away to Digi at launch remember? Where did it go?



remember this? From a person who doesn't care about power, he likes it when he pushes the power limit to crazy levels. 500+ watts with a small vcore increase under water?

hmm he even says cut down cards are heavily down clocked. HBM vs GDDR why AMD had to use HBM and if they GDDR what would have happened to Vega's power consumption.... yeah I mean come on man, we all know how much more power hungry Vega, Polaris is vs their competition. There is no denying GCN has gotten worse over the years with power consumption. You can't compare across nodes and say look its not the architecture without factoring out the node savings. If we do that its easy to see, since GCN 1.1 to now, there has been pretty much no change to GCN from a architecture stand point about power savings.

 

[_mockingbird]

You just pulled a dirty tactic by posting a picture of a benchmark without the specification which driver setting was used because it might as well been overclocked and the performance instead of default setting or power saving . And you call me straw man .

There's a name of the website at the bottom of the image. I know it's terribly exhausting to go to the website, but you have to do some work.

Never can you quote a message where I say that the power usage is good either.

Why don't you link any post of me where I say these things. Instead you post a picture of something that only proves that RX Vega is not power efficient which no one denies.

For the record I'm not asking you to explain what I already know , I'm asking to explain how this is tied to GCN (I'm not asking you to show me a pack of butter I'm asking you to show me how the butter is made and explain the process...)

You are shifting the goal post.

Earlier, you asked, "Where is the proof that each GCN design only goal is tied to using more power?".

So, now you have it.