AMD CPU speculation... and expert conjecture

[jdwii]

Still waiting on that nvidia quote Juan since no one else can find it.
edit i think we can all assume he made it up or twisted their words.
edit again
Notice how juan ignores it????????? Nvidia never said this he needs to be ban over lying!!!!

Juan toms would of banned you back in the day over that quote sometimes Nvidia engineers might actually come here back in the day we had someone working at Intel on these forums and toms mods made someone change their signature over it offending that intel worker.

I have a better proposal. Just report my signature to mods. And if some mod contact with me for asking me details about my signature I will give not only the full sources for the quotes in my signature, but I will also fill a formal petition for banning you and others for doing false accusations and continuous trolling me. Are you ready?

Done, see if anything actually happens, i think you are misinformed about several different terms one being trolling. I shown you a picture of your statement on tomshardware and i use your own statements to come to conclusions about your statements not made up crap(which would be trolling).
If however the mods agree to ban me and others i for one would support that decision as long as you were ban to at least from this thread.
Also as 8350rocks just said 1 person isn't always representative to a whole company.
However i don't want you ban to be honest i would rather have you just cite were you get your information and stop acting so egotistical and arrogant.

 

[jdwii]

[

Spoiler

quotemsg=14234312,0,1281705]

Spoiler

Nvidia has done, hilariously, what we never thought they would do and support the VESA standard for adaptive sync. http://wccftech.com/nvidia-promises-support-freesync/ I can only think of all the posts made here along the lines of "that will never happen".

If GTA5 support MANTLE then i see nVidia also supporting it, that game is just too huge.

What does that even mean? And again, why would NVIDIA care? That logic is like "Game X supports Gameworks, so AMD should adopt NVIDIA's tools because Game X is too huge".

Nevermind GTA games on the PC run like crap anyways.

Sorry to butt in, but having playing GTAIV (not the best coded GTA among the others) on maximum settings @ 1080p (view distance high, not very high), I've gotten over 80 FPS in most situations except when there are about 30 helis spawned and you are blowing them all up at the same time (lol).

This was with a GTX 770 and a overclocked i5-3570k, so it's really not that bad of a game (compared to ArmA especially). However, considering that GTAIV can barely get higher than 30 FPS in most situations @ 720p on a PS3 where as GT5 sat at 30+ FPS the whole time @ 1080p, that really shows how bad the engine was (at least for the PS3 processor / board / aluminum sheet-thing).

Also haven't been on here in a while. Any chatter about the 900 series?

Well i'm playing that game right now(beating it one last time before 5) and i get 30FPS in some cases but usually its 50-60- 90-100 in small area's. This is with very high settings or high and 30 or so on detail and view distance.

http://www.tomshardware.com/reviews/build-budget-gaming-pc,3943-18.html
I'm just not to sure if i would want to pair that CPU with a high-end GPU like that games in a few years are just going to be unplayable. I stick to my old statement for a budget 400$ build but for 500$ i would probably pick a 6300+265. GTA5 will kill this dual core Pentium.

Well we can certainly hope for that it seems like Rockstar is spending a lot of time on GTA5 for the PC since some code way back was found on the 360 disc that was for the PC.
Being such a heavy CPU title i think having Mantle might have a big influence here. Hoping it uses all 8 cores on the CPU, since the consoles have 8 cores and as stated GTA is always heavy on your CPU. I do remember i played that game 1 month ago on the 360 and the pop in was so bad it messed up my gameplay not to mention you really feel the low frame rate and i'm a person who can deal with 30fps most of the time.

 

[colinp]

Just testing my new sig...

 

[gamerk316]

http://www.gamerheadlines.com/2014/10/assassins-creed-unity-ps4-xbo-parity/
Gamer nailed this one 100-150 pages back
“We decided to lock them at the same specs to avoid all the debates and stuff. Technically, we’re CPU-bound. The GPUs are really powerful, obviously the graphics look pretty good, but it’s the CPU that has to process the AI, the number of NPCs we have on screen, all these systems running in parallel.”

“We were quickly bottlenecked by that and it was a bit frustrating, because we thought that this was going to be a tenfold improvement over everything AI-wise, and we realised it was going to be pretty hard. It’s not the number of polygons that affect the framerate. We could be running at 100fps if it was just graphics, but because of AI, we’re still limited to 30 frames per second.”

It's telling they expected so large an increase, but at the end of the day, the CPUs in the current generation of consoles, while an improvement, isn't a leaps-and-bounds improvement over what was in the 360/PS3. And it's going to be a limiting factor going forward.

 

[8350rocks]

Juan, you will notice new x86 offerings across all server scales coming soon, in addition to the ARM offerings. They are just, as of yet, unannounced.

 

[Cazalan]

This is saying Beema/Mullins are getting canned for Carrizo-L in December.

http://www.digitimes.com/news/a20141007PD202.html

I thought the cat cores were doing ok, but I guess not.

 

[Cazalan]

It's telling they expected so large an increase, but at the end of the day, the CPUs in the current generation of consoles, while an improvement, isn't a leaps-and-bounds improvement over what was in the 360/PS3. And it's going to be a limiting factor going forward.

Not much more they could do with the manufacturing capabilities at the time. The visual quality of the games are still much better than before. PS3 looked like utter crap to me after gaming on PC for so long. PS4 actually looks quite good, but definitely not up to par with my gaming rig. Gladly some key titles will be coming to PC with much higher fidelity.

 

[blackkstar]

By the way, Juan, your AMD related quote in your sig needs a caveat.

Per people I know @ AMD:

"The context of that quote was in microservers specifically, while we would love to see ARM take off in the desktop/consumer space, we have no grand illusions of such things. This is primarily due to so many massive software hurdles and compatibility issues that would stem from porting software to a new, relatively unvetted, hardware environment that x86 apps and programs are not designed for. I do hope that poor soul is not seriously running around with the mindset that x86 is ever going away. That would be unfortunate..."

The same 'people' that has been giving you massively wrong info during years?

The same people that said you then that Seattle "is not meant to even replace Opteron-X offerings, but more as an option for some who want ARM", when Seattle has replaced Opteron-X whereas AMD claims that it was because Seattle offers more performance and less power consumption? Not only the Opteron-X has been replaced but the roadmaps for 2015--2016 show ARM processors replacing the Piledriver-based Opteron 4000/6000 series, as I said then it would happen.

The same people that said you "In fact, the ARM server SoC is an experiment, that's why no custom cores, that's why very little effort into adding just the barebones necessary for a server platform. That way, if (or more like when) it bombs, they're not out retarded figures worth of research dollars, and Opteron-X will still be there sitting on the back burner. If it takes off like a rocket, great, they make money. If it doesn't (like every other ARM SoC for enterprises out there)...then they're not out massive dollars for the research and development of a custom processor" just weeks before AMD announced Skybridge project? The same people that was unaware that Keller and his team have been working in a custom processor K12 since the year 2012? Then I wrote: "AMD has already projects about custom ARM cores. The reason why are releasing Seattle with standard cores is because custom cores take time and money and AMD expect to have their server product first, before competitors release their custom chips. Sure you hear the "who strikes first strikes twice."

Is not funny that I already knew that Keller was working in a custom ARM core, when your friends at AMD didn't know it?

The same people that said you that AMD was not really interested in ARM but was only a strategic move to please HSA members? Your own words "From a diplomatic standpoint, with many ARM partners on the HSA group as founders with AMD, they would need to show a sort of support to those partners who have helped them to write the HSA standards. Now, this is exactly what this looks like..."See, ARM can work in servers!! *mumbling under breath* how long do we have to do this Rory?? I need to get back to work on steamroller...""?

Stop spreading more missinformation. Two years are enough.

Juan, let me help you understand why your logic is faulty. Lets hop back in a time machine to the mid Pentium 4 era.

Do you remember what Intel's roadmaps looked like? They had 10ghz Netburst on it. That was their grand plan.

Now envision people on forums talking about 10ghz Netburst coming in a few years. That is what was going to happen and Intel didn't know any better about future plans. Roadmaps, Intel employees, insiders, they all talked about how Netburst was going to scale to great frequencies.

Eventually, that 10ghz netburst never showed up. Because it got cancelled due to technical reasons. Reasons that were beyond Intel's control. Yet at a time in history, Intel constantly spoke of high frequency single cores designs.

The proper thing to do is to blame Intel and the laws of physics for failing us. However, what you are doing is basically attacking the messenger and going "well you guys said that we'd see 10ghz Netburst and we never did so you're all wrong! Everything you say is wrong because you've been wrong before!"

Does Intel changing their plans mean all the people who spoke of what Intel was doing was wrong? Were they all misinformed fools?

NO! They just were messengers of what Intel was planning on doing!

Do you see the error in your logic? You're shooting the messengers when the message they received was wrong in the first place. The reason why people are going after you, is because you are not a messenger. You are making explicit claims about directions of products as well as performance. And you're extrapolating these and drawing these conclusions on your own. So you're in a much different position than people like FX who are just messengers. You are a creator of a message and a messenger at the same time.

Hopefully this makes things a little more clear to why you get treated like you do on the forums.

 

[Saga Lout]

The next poster to call another a liar gets a month's ban out of here - and that ain't no lie!.

 

[jdwii]

Juan it has nothing to do with me it has to do with reality "extraordinary claims need extraordinary evidence"
Again i don't want you ban i want you to cite your information and not claim someone claimed something without evidence
Edit again i did nothing i only showed your own statements on this site would you like me to do it again???? I have to say it again i don't want you ban i want you to show and prove your own statements not lie its as simple as that. Mods will not except your BS when you don't give evidence to support your claims when they are major.
Edit again i don't like anyone's statements when its based on lies or fallacious material i stated this several times if you feel that includes your own statements then fine

Nope. For several days you required mods to ban me if I cannot give a source for my signature. I will provide the source when they contact me. At same time it is fair that your own logic is applied to yourself and you are banned when you cannot provide sources for the several claims that you been making in this thread, because your "my one and only contact with a friend who works at Amd told me" and your "this was confirmed from my friend at Amd" are not valid sources according to the strict standards that you want apply to others in this thread. Thus better you prepare to be banned soon.

i don't even have that person anymore to get info from he isn't working at Amd anymore. Not to mention that was 120 pages ago at least

 

[gamerk316]

More benchmarks of ARM servers compared to Xeon. I recall when some people in this thread predicted that ARM servers never never never could caught x86 in performance. Let us see

The Haswell server is using the octocore E5-2630v3. The Ivy Bridge server is using the sexacore E5-2630v2. The Sandy Bridge server is using the sexacore E5-2630. Note that the Intel chips have hyperthreading. Thus they are 16, 12, and 12 threads processors respectively, whereas the XGene processors only execute 8 threads.

On the light of this new data I maintain my former claim that AMD change of strategy to conquer the server market is essentially correct. I continue maintaining that AMD's K12 core will be at Haswell level of performance in scalar/integer workloads and at Ivy Bridge level of performance in vector/FP workloads.⁽¹⁾ However the efficiency will be superior to x86, which means AMD will be able to bring "moar cores" in the same power consumption envelope. I expect that the K12 servers will be faster than Broadwell/Skylake servers consuming about the same performance. How much? It is still very difficult to say.

⁽¹⁾ Disclaimer: I can change my opinion at any time that I can obtain more data about the architecture.

What workload? Because I get the sense compute is coming into play here. Course, plug in a Tesla and use that instead, you get different performance characteristics. So not having a benchmark title is a giant red warning light that the marketers are making the same mistakes all over again.

 

[8350rocks]

The next poster to call another a liar gets a month's ban out of here - and that ain't no lie!.

LOL @ your sig...

hahahaha

 

[-Fran-]

You are ommiting several important details that when are considered show that your reasoning is flawed.

In essence some people is making bogus claims, I am correcting them and their reaction to my corrections has consisted on:

(i) Try to hide their mistakes or try to attribute them to me.

(ii) Try to negate the correctness of my predictions at any cost.

(iii) Critizicing what I never said, but omitting that I really said.

(iv) Attempts to attack my credibility claiming that I know nothing despite many of my claims/predictions have been proven lots ot times during years.

(v) And more.

Several posters have made a clever resume of what is happening in this thread. I recommend you to read some of their posts, for instance the next one

http://www.tomshardware.co.uk/forum/352312-28-steamroller-speculation-expert-conjecture/page-325#14320911

Juan, his point is that all of your predictions, independent of being correct or not, are from what the companies themselves say in one way or another (roadmaps, interviews, internal information, etc). His point is that if the company makes a mistake, you make a mistake by association. And that applies to everyone, not just you. What he adds, is that you make additional calculation based on those "projections", meaning you'll be pretty close (depending on your math) or horribly wrong. But you are not accepting the first (and most important) part: a company can be horribly wrong when projecting or predicting the future. AMD is really way off track at times and we don't need to force memory for mistakes they've made. That's where you're not accepting arguments back and defend yourself only saying "the engies know better". No, they also make mistakes. That's the important bit in blackkstar's argument (which was pretty good, by the way).

Cheers!

 

[GOM3RPLY3R]

@ juan

Your baseline is not 0. Get that figred out.

You didnt guess 6000 off of 0.

http://www.tomshardware.co.uk/forum/352312-28-steamroller-speculation-expert-conjecture/page-326#14327681

^^^^ What part of the definition of % error and how is computed to the predicted value of 6000 you didn't understand?

I have to go with juan here on a mathematical sense:

(5681) - (6000)
-------------------- X 100 = 5.316%
(6000)

(5681) - (6000) = -319 / (6000) = -0.05316... X 100 = -5.316% (no negative percents in percent error) = 5.316%



Though I'm not on it that much, I used to love this thread because of the debate, but now most people (not all) are just a bickering class of 8th graders. I am still young (16), and I used to be a bit immature on this thread and shouted stupid replies and argued moot points. However now that I like to debate things that I know about and can debate about, it sucks to see most people just yelling at each other getting nowhere in the conversation.

The point of a thread like this, and as any debate usually goes, is to share opinions and to give valid arguments to your opinions, not to call people things, accuse them of stuff, and even if you have evidence, to prove them wrong.

What good is a game made from JavaScript or C# if you can't get your arguments right? It just leads to something broken and glitchy, and no one would want to play it. However if you do get your arguments right, it could be a great game and everyone would want to play it.

 

[8350rocks]

@yuka: precisely.

 

[szatkus]

Thread has gone on really wrong way.

From time to time I had here some great discussions with you, guys. Except one person...

 

[Cazalan]

More benchmarks of ARM servers compared to Xeon. I recall when some people in this thread predicted that ARM servers never never never could caught x86 in performance. Let us see

Who said they wouldn't catch x86 in any benchmark whatsoever? The question was would they scale AND keep a power advantage. In this particular unnamed benchmark they're tied, so the ARM power advantage just disappeared. Keep unreleased product out of it, as even the X-Gene1 product is hard to get hold of today.

You're presenting this as a good thing for AMD, but it's really just highlighting that they're about 2 years behind AMCC. If a big player like Samsung or Qualcomm wants to get serious about ARM servers they can snap up AMCC for 1/8th the price of AMD. In which case AMD is in a rough spot regarding their ARM ambitions.

 

[8350rocks]

More benchmarks of ARM servers compared to Xeon. I recall when some people in this thread predicted that ARM servers never never never could caught x86 in performance. Let us see

Who said they wouldn't catch x86 in any benchmark whatsoever? The question was would they scale AND keep a power advantage. In this particular unnamed benchmark they're tied, so the ARM power advantage just disappeared. Keep unreleased product out of it, as even the X-Gene1 product is hard to get hold of today.

You're presenting this as a good thing for AMD, but it's really just highlighting that they're about 2 years behind AMCC. If a big player like Samsung or Qualcomm wants to get serious about ARM servers they can snap up AMCC for 1/8th the price of AMD. In which case AMD is in a rough spot regarding their ARM ambitions.

Another reason that they are not massively hyping all of this outwardly.

As an established x86 player, they have clout in those markets, as an unestablished ARM player, they have zero clout that Samsung, Motorola, AMCC and others have spent years trying to build up.

Now, a player the size of AMD coming into that market validates those companies in many ways that the x86 market would otherwise have largely ignored for much longer in all likelihood. However, it likely benefits those companies that AMD entered those markets far more than it benefits AMD.

Think of it like this...AMD/Intel are like the U.S. in NATO...ARM is like a new emerging country, and they have just been recognized as a country by the U.S. Which means that now they will start to get a lot of the attention and support that your average 3rd world country would get.

 

[Cazalan]

Lol you must have been reading different articles cause the ones I saw said 5.5Ghz air/water from the beginning. It's right in the HWBot rules.

http://hwbot.org/competition/intel_occ_computex14/

 

[jdwii]

@ juan

Your baseline is not 0. Get that figred out.

You didnt guess 6000 off of 0.

http://www.tomshardware.co.uk/forum/352312-28-steamroller-speculation-expert-conjecture/page-326#14327681

^^^^ What part of the definition of % error and how the error is computed using the predicted value of 6000 and the measured value of 5681 you didn't understand?

Juan it has nothing to do with me it has to do with reality "extraordinary claims need extraordinary evidence"
Again i don't want you ban i want you to cite your information and not claim someone claimed something without evidence
Edit again i did nothing i only showed your own statements on this site would you like me to do it again???? I have to say it again i don't want you ban i want you to show and prove your own statements not lie its as simple as that. Mods will not except your BS when you don't give evidence to support your claims when they are major.
Edit again i don't like anyone's statements when its based on lies or fallacious material i stated this several times if you feel that includes your own statements then fine

Nope. For several days you required mods to ban me if I cannot give a source for my signature. I will provide the source when they contact me. At same time it is fair that your own logic is applied to yourself and you are banned when you cannot provide sources for the several claims that you been making in this thread, because your "my one and only contact with a friend who works at Amd told me" and your "this was confirmed from my friend at Amd" are not valid sources according to the strict standards that you want apply to others in this thread. Thus better you prepare to be banned soon.

i don't even have that person anymore to get info from he isn't working at Amd anymore. Not to mention that was 120 pages ago at least

Irrelevant. My discussion about future 2020 APUs was also hundred of pages ago. I will continue applying to you the same logic/standards that you apply to me.

I suppose that is fair lets just be done with this hate now its really not a drama thread its a thread about future speculation on CPU's i find your input to be quite useful for some.

 

[GOM3RPLY3R]

@ juan

Your baseline is not 0. Get that figred out.

You didnt guess 6000 off of 0.

http://www.tomshardware.co.uk/forum/352312-28-steamroller-speculation-expert-conjecture/page-326#14327681

^^^^ What part of the definition of % error and how is computed to the predicted value of 6000 you didn't understand?

I have to go with juan here on a mathematical sense:

(5681) - (6000)
-------------------- X 100 = 5.316%
(6000)

(5681) - (6000) = -319 / (6000) = -0.05316... X 100 = -5.316% (no negative percents in percent error) = 5.316%

I don't know if this is known here, but the formula that appears in my post or in the Wikipedia link uses vertical bars | |.

| (5681 - 6000) / 6000 | * 100

Those vertical bars denote absolute value

|−0.05316666| = 0.05316666

This is why at the end we obtain a positive value of 5% using the formula for the % error.

Never learned that in school, but that makes a ton more sense now that I look at it, lol. My teachers always said, "Just make it positive cause it has to be." Thanks for the correction.

Edit: Also about the quote, "Discrete GPUs will be killed by about 2020," I recall that after watching either the 2013 or 2014 Nvidia conference that Jen-Hsun Huang had said that. I may be wrong, and I tried looking for it myself. Though I have not found it yet, I did only look for about 15 minutes before taking a break. I may be wrong, but I just wanted to put it out there. It may explain why it is not found in text, but possibly in voice.

 

[griptwister]

@GOM3RPLY3R, Your age explains so much! Haha, thats cool you got into PCs at a young age.

@everybody else, I've kept silent for so long cause honestly, I got bored with rummaging through these posts looking for something of value and end up seeing useless arguments. The truth is Juan seems to spit out information based on hypothesis to draw attention to his website, "Juan's Hardware." :lol:

In all seriousness, it's really irritating seeing you guys argue with him instead of ignore him. The more you guys talk, the more it makes what he's doing, acceptable.

 

[palladin9479]

Problem solved. Cease all disruptive and negative behaviors. No more personal attacks or extremely aggressive posts. Lets all be adults, and if your not capable of doing that then you will be removed.

 

[GOM3RPLY3R]

@GOM3RPLY3R, Your age explains so much! Haha, thats cool you got into PCs at a young age.

Haha, thanks.

Just as a side note, does anyone have, have had experience with, or have been following the GTX 900 series, more precisely the GTX 970 itself? I have some information to share however I want to make sure it's able to be built upon in a structural and mature manner. ^_^

 

[jdwii]

Problem solved. Cease all disruptive and negative behaviors. No more personal attacks or extremely aggressive posts. Lets all be adults, and if your not capable of doing that then you will be removed.

Thank you and i will try my best for now on sorry bro