AMD Financial Analyst Day (Official Zen & GCN News)

[patrickjp93]

There's a difference between PR and internal numbers. People who fail to realize that haven't had enough experience in these types of situations. There are no reviews for Zen (for obvious reasons) so unless you have a source proving otherwise your doubts are immediately extinguished. For the time being 40% stands until reviews come later backing it up. I'm sure we'll see demos of the product long before launch given such a drastic performance increase is something that will later make for good PR. No one is forcing you to believe it although its not a number that's entirely out of reach for a company hasn't been wrong with their past three architectures. There's no advantages for falsifying performance figures.

Yes there is: hype for sales.

[patrickjp93]

There's no advantage to falsifying the information of your own product. Using an analogy much like yours if you were a math teacher that made a quiz for your students. Would you grade them based on what they think are the right answers? Certainly not. Nobody knows more about a product or item than its creator. A prime example would be my LinusTechTips Tray Tool. You wouldn't even know where to start with explaining how it works and whether or not to question any performance improvements I do to the software as you know I could run full circle around you with it. Why? Because I am the original creator nobody knows how it works better than I do. Zen was a from scratch architecture designed specifically by Keller and his team. Expecting anything less than 40% over a Bulldozer revision would immediately throw fanboy flags. This isn't a debate whether or not it will deliver the full 40% but the fact that the 40% is set and stone right now as it cannot be debunked by any theory. All we have now is a couple casuals coming into the thread touting "omg 40% no wai too mach AMD can't do it" not so cleverly disguised as biased opinions.

 

There's a distinct difference between the PR I am referencing and internal testing. Coming out and saying Zen will have 40% higher IPC over Excavator wasn't a publicity stunt but more over of what to expect from the upcoming product (keep in mind this is Financial Analysts Day). Once again you're questioning the creator, something that will never hold any weight. You will see how far that will carry you in this thread as you two are the only ones who honestly think that AMD is just throwing bullshit into the fan. There has to be a logical explanation for that somewhere...

 

Mr Moose's posts imply what I've stated in the above paragraph so I think a lot more than just me would like to see credible proof as he flat out stated that AMD's numbers are more than likely wrong. Not true, numbers can be dead accurate without any evidence backing them up. Lets go back to the software I wrote for this community for another example. If I posted in the thread that I was going to integrate the entire forum such as the ability to reply to messages, posts, etc straight from the tool does that mean the information I shared with you guys is wrong? Simply because I haven't launched a revision of the software with them features? Even though I may already have an internal build with every one of them features working undergoing extensive testing. You see there's backwards logic and reality. Some of us live in the real world while others live in their own fantasy world. The door swings both ways, you simply cannot prove the 40% wrong either. Given the company was 100% accurate with their past three architectures I think it's safe to say it's perfectly fine to speculate based off that number.

 

I don't full on trust the numbers as several of my posts in this thread clarifies. I just believe Zen should deliver around that area of performance given the circumstances and I don't mind speculating and building information based off that data. It's the only bridge to stand on right now as that's official word without any other information coming from anywhere else (it's the only real information thus far). If you think it's fake, fine that's all you. If you take it to heart 100% and live the next year and half of your life by that number, more power to you. Although you simply cannot just insist that it's fake based off no information at all.

That's not how marketing or consumer thinking works. You think too much like an enthusiast and not nearly enough like the average dumb consumer or Company X fanchild.

[Victorious Secret]

Why are people getting so butthurt that no one is taking a marketing push at its word? Are we stupid for not wanting independent testing to verify the claims? People like to bitch about how Nvidia lied with the 970, isn't how we found out by way of INDEPENDENT TESTING? 

Why is it that Intel or Nvidia can be harassed but AMDs shit can never stink? Are some of you that naive? 

We wait for a 3rd party to test these claims, like they always do. THEN they become a fact. Until then it is AMD pumping their tires for their share prices; not for your benefit. 

 

Damn, why do people so willingly eat up what one company says yet get angry over what another says? Treat all of them the same. Stop being such fanboys and stop getting offended when other people ask why you're so personally offended when people don't "toe the party line" 

[Sakkura]

Why are people getting so butthurt that no one is taking a marketing push at its word?

 

Why are people getting so butthurt about the AMD numbers, and even the slightest hint of anyone taking them seriously?

 

Corporations don't always lie.

[zappian]

Anyone remember this?

 

http://www.extremetech.com/gaming/180088-nvidias-questionable-geforce-337-50-driver-or-why-you-shouldnt-trust-manufacturer-provided-numbers

 

We'll see what actual tests bring, rather than go by the first-party numbers.

 

40% over excavator is still like ivy bridge IPC .

Why is that so hard to believe . Thats not that amazing at all , its still meh IPC compared to skylake.

I swear some people on this forum just want to see AMD sink and fall so they can pay more for their intel processors ( monopoly).

They didnt even throw out impossible numbers , its still under what intel has NOW.

Very reasonable numbers .

This forum is so green and blue sometimes *sigh*

[LAwLz]

40% over excavator is still like ivy bridge IPC .

Why is that so hard to believe . Thats not that amazing at all , its still meh IPC compared to skylake.

I swear some people on this forum just want to see AMD sink and fall so they can pay more for their intel processors ( monopoly).

They didnt even throw out impossible numbers , its still under what intel has NOW.

Very reasonable numbers .

This forum is so green and blue sometimes *sigh*

Again, I don't see anyone here claiming that AMD can't manage a 40% increase. I am just saying that you shouldn't trust the number because they haven't been validated yet.

You should never trust performance metrics that come straight from the company itself. Always wait for independent tests before making your mind up about a product, no matter what company is behind it.

 

 

I think this forum needs to tune down the "if you aren't 100% loyal to company X then you are a hater and must like company Y instead". There are many shades of yellow between red and blue/green.

[Glenwing]

40% over excavator is still like ivy bridge IPC .

Why is that so hard to believe . Thats not that amazing at all , its still meh IPC compared to skylake.

I swear some people on this forum just want to see AMD sink and fall so they can pay more for their intel processors ( monopoly).

They didnt even throw out impossible numbers , its still under what intel has NOW.

Very reasonable numbers .

This forum is so green and blue sometimes *sigh*

 

It's not the number that's hard to believe, it's the source. Based purely on past experience, we should be skeptical of first-party claims about important high-profile products until we see it verified independently.

[shawno]

I think people are still hurting from amds bulldozer debacle. Before it was released they gave it massive praise but when released it wasnt anything to write home about.

However: 40 percent increase in performance is almost as good as nvidia stating that pascal would be 10x faster than maxwell.

Ceo math is amazing. Lets hope theyre both right.

[mr moose]

Why are people getting so butthurt about the AMD numbers, and even the slightest hint of anyone taking them seriously?

 

Corporations don't always lie.

 

 

40% over excavator is still like ivy bridge IPC .

Why is that so hard to believe . Thats not that amazing at all , its still meh IPC compared to skylake.

I swear some people on this forum just want to see AMD sink and fall so they can pay more for their intel processors ( monopoly).

They didnt even throw out impossible numbers , its still under what intel has NOW.

Very reasonable numbers .

This forum is so green and blue sometimes *sigh*

 

No one is finding hard to believe or are buthurt over "numbers",  we're just not gullible fools who take it at their word and we will wait for third party testing.

 

To be honest I don't know what's worse,  the fact that this is such a simple concept yet it seems to be too hard for some to understand (or a too fanboy about it) or the fact that people think we are somehow lacking to want third party testing and not trust first party claims. 

[patrickjp93]

40% over excavator is still like ivy bridge IPC .

Why is that so hard to believe . Thats not that amazing at all , its still meh IPC compared to skylake.

I swear some people on this forum just want to see AMD sink and fall so they can pay more for their intel processors ( monopoly).

They didnt even throw out impossible numbers , its still under what intel has NOW.

Very reasonable numbers .

This forum is so green and blue sometimes *sigh*

It's actually almost 5% faster than Haswell by my calculations.

[Victorious Secret]

No one is finding hard to believe or are buthurt over "numbers", we're just not gullible fools who take it at their word and we will wait for third party testing.

To be honest I don't know what's worse, the fact that this is such a simple concept yet it seems to be too hard for some to understand (or a too fanboy about it) or the fact that people think we are somehow lacking to want third party testing and not trust first party claims.

Do you trust car manufacturers on their safety results? Fuck no.

You trust the independent agency who has the sole job of crashing the cars and reporting the unbiased data.

that seems like too hard a concept for people here to understand.

[Briggsy]

I think people are still hurting from amds bulldozer debacle. Before it was released they gave it massive praise but when released it wasnt anything to write home about.

However: 40 percent increase in performance is almost as good as nvidia stating that pascal would be 10x faster than maxwell.

Ceo math is amazing. Lets hope theyre both right.

 

. I bet no matter how much AMD try to hype up Zen, general expectations are going to remain low until launch. AMD chose to fart in our general direction for 8 years while they went off on a magic-filled adventure to Integer-Land. Now they are back from their bender telling us they have tiger blood, and that they are winning. I hope they pull off another grand slam in the CPU market, but like you say CEO math is amazing, and no one in the tech review industry wants to look like a fool again.

[MageTank]

. I bet no matter how much AMD try to hype up Zen, general expectations are going to remain low until launch. AMD chose to fart in our general direction for 8 years while they went off on a magic-filled adventure to Integer-Land. Now they are back from their bender telling us they have tiger blood, and that they are winning. I hope they pull off another grand slam in the CPU market, but like you say CEO math is amazing, and no one in the tech review industry wants to look like a fool again.

 

Problem is AMD seems like they are putting all their eggs in one basket this time. Bringing Keller on board is smart, but the man is no wizard. Even if you redo the entire architecture from the ground up, and get back on track, what exactly can AMD do that Intel cannot at this point? Intel has vast resources and are catching up in pretty much every front in computing. If Iris Pro is as good as it is rumored to be, then even AMD's APU's are in trouble. There is that thread in which AMD was quoted as being "tired of being referred to as the cheaper option" but thats pretty much all AMD has going for them.

 

If AMD thinks they can tough the enthusiast market with Zen, they better pull out all stops. We need feature heavy boards, and performance comparable to Intel's offerings. Perhaps expanding to other area's would help too. ITX boards would definitely be a nice change of pace for the AMD side of things.

 

AMD's solution of throwing more cores at a problem is only detrimental to their success, that is until software finally catches up. Until then, AMD is only shooting themselves in the foot.

[Victorious Secret]

-snip-

 

Thank you for being a part of the problem that we are speaking out against. Congrats, you're using something made decades ago as a standard to which their current products are made? Really? Them making seatbelts has FUCK ALL to do with their 2015 crash test results. Nothing. 

 

You shouldn't trust any companies statements. Thats what we are saying. It takes independent and unbiased testing to confirm those statements. Just cause Volvo passed last years model doesn't mean this years model will too. It goes for every manufacturer. 

Calling me an idiot? Thats rich. The only person blindly following anothers word is you and others like you in this thread. 

[Sakkura]

what exactly can AMD do that Intel cannot at this point?

 

They could theoretically bet on the high-performance market that Intel has abandoned in order to focus on mobile. You know why the architectures after Sandy Bridge have been so disappointing on the desktop? Because Intel has optimized their architectures and process nodes for much lower power consumption, at the expense of clock speeds at the high end. To get a Core M, you sacrifice 5+ GHz Core i5/i7.

[MageTank]

They could theoretically bet on the high-performance market that Intel has abandoned in order to focus on mobile. You know why the architectures after Sandy Bridge have been so disappointing on the desktop? Because Intel has optimized their architectures and process nodes for much lower power consumption, at the expense of clock speeds at the high end. To get a Core M, you sacrifice 5+ GHz Core i5/i7.

 

Yes, but it is a steady improvement nonetheless. Looking at the Thuban, compared to the first set of Bulldozers, AMD actually brought less performance with that architecture change. You are correct in saying that they could do it, but their history has turned most of the PC world into skeptics. This is why the claim of 40% IPC improvement is being shrugged off as fairy dust and snake oil. At this point, it just seems too good to be true. 

 

@Victorious Secret: I do not think Sakkura was calling you an idiot per say, he was saying that anyone in general not viewing volvo's history of setting safety records and standards and not believing they are safe, is an idiot. To his credit, volvo has built a global reputation of being very safety-oriented when it comes to their vehicles. Even those of us in the US recognize the work they put into safety. Their history speaks for them even to this very day. They improved upon original Seat Belt designs, including inventing the 3-point system that they then gave to the entire world to use openly in the name of safety.

 

While i agree that trusting a companies word blindly is a terrible way to go about things, if a company has a prominent history on delivering what they say they can do, then faith in their ability can be understood.

[mr moose]

Yes, but it is a steady improvement nonetheless. Looking at the Thuban, compared to the first set of Bulldozers, AMD actually brought less performance with that architecture change. You are correct in saying that they could do it, but their history has turned most of the PC world into skeptics. This is why the claim of 40% IPC improvement is being shrugged off as fairy dust and snake oil. At this point, it just seems too good to be true. 

 

@Victorious Secret: I do not think Sakkura was calling you an idiot per say, he was saying that anyone in general not viewing volvo's history of setting safety records and standards and not believing they are safe, is an idiot. To his credit, volvo has built a global reputation of being very safety-oriented when it comes to their vehicles. Even those of us in the US recognize the work they put into safety. Their history speaks for them even to this very day. They improved upon original Seat Belt designs, including inventing the 3-point system that they then gave to the entire world to use openly in the name of safety.

 

While i agree that trusting a companies word blindly is a terrible way to go about things, if a company has a prominent history on delivering what they say they can do, then faith in their ability can be understood.

 

The problem here is the concept that taking a companies word for it is naive. Shakkura is condoning a naive approach to accepting performance data. He is using one companies safety record throughout all of it's history as a reason to blindly believe every other company at their word.

 

This, however, is not what Victorious said or even implied.  It was quite clear he was talking about the industry as a whole and as pointed out even if you have a company that repeatedly makes a safe car or secure OS or even a robust laptop, there is no guarantee the next one will be. 

 

There is no excuse for believing what a company says about it's own product.  We don't do it with nvidia, we don't do it with Intel, we don't do it with Asus or any of the tech companies.  In fact rational people don't do it for any company.  So the only person who would do it would be someone who is irrational, someone who is naive or someone who is a blind fanboy.  PERIOD.

[MageTank]

The problem here is the concept that taking a companies word for it is naive. Shakkura is condoning a naive approach to accepting performance data. He is using one companies safety record throughout all of it's history as a reason to blindly believe every other company at their word.

 

This, however, is not what Victorious said or even implied.  It was quite clear he was talking about the industry as a whole and as pointed out even if you have a company that repeatedly makes a safe car or secure OS or even a robust laptop, there is no guarantee the next one will be. 

 

There is no excuse for believing what a company says about it's own product.  We don't do it with nvidia, we don't do it with Intel, we don't do it with Asus or any of the tech companies.  In fact rational people don't do it for any company.  So the only person who would do it would be someone who is irrational, someone who is naive or someone who is a blind fanboy.  PERIOD.

 

This argument can be made either way, you understand that, right? History has shown us that the sun rises each day, are you under the impression that the sun has not risen until you have personally seen it? Making an assumption on something based on previous history is not a far-fetched notion. Yes, it is naive to automatically assume what is being told to us to be 100% fact, but if their history has complied with their claims every single time, then it is probably safe to at least give them the benefit of the doubt. 

 

Victorious used cars as an analogy, and Sakkura brought up a point involving one manufacturers history. Did he go about it in the nicest way? Probably not, but I understood his point. While i personally do not believe AMD in this regard (because of their history) i think it is ignorant to say that one cannot take the history of a company into consideration before weighing in on whether or not their claims hold merit. If Volvo tells me they made an improvement on their safety features, i will believe them until they are proven wrong, because history dictates that they have been successful with this specific detail more times than not.

 

Call me naive, but that's how it works with many aspects in life, just not quite the same for computer hardware.

[mr moose]

This argument can be made either way, you understand that, right? History has shown us that the sun rises each day, are you under the impression that the sun has not risen until you have personally seen it? Making an assumption on something based on previous history is not a far-fetched notion. Yes, it is naive to automatically assume what is being told to us to be 100% fact, but if their history has complied with their claims every single time, then it is probably safe to at least give them the benefit of the doubt. 

 

Victorious used cars as an analogy, and Sakkura brought up a point involving one manufacturers history. Did he go about it in the nicest way? Probably not, but I understood his point. While i personally do not believe AMD in this regard (because of their history) i think it is ignorant to say that one cannot take the history of a company into consideration before weighing in on whether or not their claims hold merit. If Volvo tells me they made an improvement on their safety features, i will believe them until they are proven wrong, because history dictates that they have been successful with this specific detail more times than not.

 

Call me naive, but that's how it works with many aspects in life, just not quite the same for computer hardware.

 

I don't see how you can reverse that argument,  We are not talking about a cycle like the earth spinning which can be easily predicted, we are talking about a struggling company run by humans that need consumers to be enthused enough not to upgrade until they can get zen out, and investors enthused enough not to sell all their shares.  There is just no reason to believe them until independent testing is done.  As I already said, I don't believe any other company, why would I believe AMD? It's simply not rational.

[MageTank]

I don't see how you can reverse that argument,  We are not talking about a cycle like the earth spinning which can be easily predicted, we are talking about a struggling company run by humans that need consumers to be enthused enough not to upgrade until they can get zen out, and investors enthused enough not to sell all their shares.  There is just no reason to believe them until independent testing is done.  As I already said, I don't believe any other company, why would I believe AMD? It's simply not rational.

 

We are talking about the notion of belief based on history. The comparison to car manufacturers was made, and when someone brings up a point regarding a company with an excellent record in doing what they say they do, you call it naive. You call cycles of the earth "easily predicted". Is this because it happens often enough to just assume it will happen again? 

 

My point is rather simple. You CAN make assumptions based on history of a company. Your assumption can work both ways too. You can assume it will do badly  based on its history, and you can also believe it will do good, based on its history. Belief in general is subjective, and differs from person to person. We are talking the natural balance of inferences and assumptions all together. 

 

I do see your point, and i do agree with your way of thinking. Your rational is based around critical thinking, but mine is too. Part of critical thinking is not just waiting to see something happen. It also applies to learning information based on previous experience and applying it to relative judgement on a subject. Companies have a history of inflating their performance, so you are inclined to believe every company can do this too, so you judge them all in that regard. I judge based on each companies individual history, under the exact same guidelines as you, but i do not pool no two companies together. Some companies really do value their word, and are truthful in what they report. If they have a solid history in doing so, we tend to take their word for it until they tarnish their own records. We use this metric when recommending parts to people, do we not? We tell people how a piece of hardware from a specific company has performed for us, and how certain brands are "safer" to buy than others? You cannot say you are not guilty of this, because every human tends to think this way about one thing or another.

 

I think this will end better if we all let each individual person make estimates on how they feel AMD will perform, and not crucify them for believing in something. I think i am gonna stop talking now, because i feel like i overused the words "believe" and "assume". 

 

TL:DR? People think stuff all the time. Let them continue thinking it using their own basis for doing so.

[mr moose]

We are talking about the notion of belief based on history. The comparison to car manufacturers was made, and when someone brings up a point regarding a company with an excellent record in doing what they say they do, you call it naive. You call cycles of the earth "easily predicted". Is this because it happens often enough to just assume it will happen again? No, the science behind it is well understood and reasoning for knowing it will happen again doesn't rest on researches needing people to buy their product.  

 

My point is rather simple. You CAN make assumptions based on history of a company. Your assumption can work both ways too. You can assume it will do badly  based on its history, and you can also believe it will do good, based on its history. Belief in general is subjective, and differs from person to person. We are talking the natural balance of inferences and assumptions all together.  You can make assumptions and you can be wrong for doing so.

 

I do see your point, and i do agree with your way of thinking. Your rational is based around critical thinking, but mine is too. Part of critical thinking is not just waiting to see something happen. It also applies to learning information based on previous experience and applying it to relative judgement on a subject. Companies have a history of inflating their performance, so you are inclined to believe every company can do this too, so you judge them all in that regard. I judge based on each companies individual history, under the exact same guidelines as you, but i do not pool no two companies together. Some companies really do value their word, and are truthful in what they report. If they have a solid history in doing so, we tend to take their word for it until they tarnish their own records. We use this metric when recommending parts to people, do we not? We tell people how a piece of hardware from a specific company has performed for us, and how certain brands are "safer" to buy than others? You cannot say you are not guilty of this, because every human tends to think this way about one thing or another.  Just different levels of acceptance here I think.

 

I think this will end better if we all let each individual person make estimates on how they feel AMD will perform, and not crucify them for believing in something. I think i am gonna stop talking now, because i feel like i overused the words "believe" and "assume". 

 

TL:DR? People think stuff all the time. Let them continue thinking it using their own basis for doing so.

 

I have no problem with people getting excited about the numbers  and believing them, but I do draw the line at outright saying they are gospel truth on forums where less educated people come for information.

 

The reality is I am really hoping they perform better than promoted.  I however won't be telling anyone anytime soon to bank on it, and as a result I will not be telling anyone to "wait for ZEN" if they want to upgrade now.

[MageTank]

I have no problem with people getting excited about the numbers  and believing them, but I do draw the line at outright saying they are gospel truth on forums where less educated people come for information.

 

The reality is I am really hoping they perform better than promoted.  I however won't be telling anyone anytime soon to bank on it, and as a result I will not be telling anyone to "wait for ZEN" if they want to upgrade now.

 

This is a fair statement i can get behind. I too hope Zen is every bit as good as it is being promoted as being, we really do need the competition as consumers. The quicker AMD gets back into an offensive stance against intel in the CPU department, the better prices and advancement will be for us, the consumers. Considering Zen is still in the dark on its official release date, and knowing it will not happen for at least another year or so, waiting on a "maybe" would be a bad idea for anyone doing so. If you need more performance now, then upgrade now. That would be my advice to those people.

 

I don't really care if AMD surpasses intel in performance, i just want to see more thermal headroom this time. These 95w/125w CPU's are making my dream of having a gaming oriented AMD ITX build a lost cause. 

[Eroda]

whether its 40% is actually irrelevant at this point, we have no other idea about any other features of Zen and the Chipsets  so whats wrong with taking AMD at their word and thinking well if thats true it might be a compelling product when it launches but until we have some more information about other features like the possibility of a fusion core being on die for all Zen APU/CPU there shouldn't be any issue with taking AMD at their word currently. 

[Humbug]

Even if you redo the entire architecture from the ground up, and get back on track, what exactly can AMD do that Intel cannot at this point? Intel has vast resources and are catching up in pretty much every front in computing.

AMD does not have to beat outright beat Intel to be profitable. They have to have a good architecture, good execution and competitive performance which allows them to sell at decent margins in decent quantities and start making profits. That's what Zen has to do. What happened with bulldozer (and it's derivatives) was that they fell so far behind that they had to sell everything really cheap with almost no margins in small quantities, and they had to temporarily give up on the high performance desktop and server markets.

[MageTank]

AMD does not have to beat outright beat Intel to be profitable. They have to have a good architecture, good execution and competitive performance which allows them to sell at decent margins in decent quantities and start making profits. That's what Zen has to do. What happened with bulldozer (and it's derivatives) was that they fell so far behind that they had to sell everything really cheap with almost no margins in small quantities, and they had to temporarily give up on the high performance desktop and server markets.

 

Bulldozer was a huge step backwards. Those cores came at a heavy cost, and that cost was AMD's gaming market. I said in a couple posts back that AMD would not have to surpass Intels performance. If they can offer comparable performance, in more than just one form factor, they could win back enough of the market to expand. They just really need to make sure its a solid design this time. If AMD makes another terrible series of CPU's, i fear it might spell an unrecoverable disaster.