AMDOMINATION

>Pushed back to 2017
>Q3 2016 brings Skylake-E

AMD sure is dominating the field of beating Intel two years too late.

AMD is just at the mercy of Intel. They don't want to kill them and trigger antitrust laws.

File: 1464625325199.png (398KB, 792x540px) Image search: [Google]

398KB, 792x540px

>>56180726
>can compete with Intel's best

And so the overhyping begins resulting in a huge letdown. It can compete with Intel's medium tier CPUs at best.

>>56180751
You're mixing up Skylake-E and Kaby Lake. You'll have to wait for Q1/Q2 2017 for the new meme platform.

>>56180769
/thread

High school teacher's friend is an engineer at Intel. They have shit in the labs 10 years ahead of what's commercially available ready to crush anything AMD puts out, and that was 10 years ago.

File: 1470608904174.jpg (35KB, 659x741px)

35KB, 659x741px

>>56180905
>being in high school

>>56180937
>being an illiterate high schooler

>>56180726
We all know it's not going to do shit.

AMD always promises this bullshit and never delivers.

AMD = Another Major Disappointment

>>56180983
>always

Nigger it's been 6 years since the last AMD CPU gen

Don't confuse AMD CPU division with AMD GPU division

Why did they downclock the Intel Core i7-6900K CPU to 3GHz instead of overclocking the AMD Zen CPU? I know why. Because the AMD Zen CPU probably won't go much above 3GHz. Meanwhile all Intel CPUs pretty much get to 4.3/4.4GHz.


AMD is dead.

>>56180726
We are having this thread for 5 years already
AMD and intel are both shit, both of them have retarded fanboys keeping their jewishim alive and well and bith of them will never release anything with more than 5% perfomance improvement (at most) every year

File: 1443494003041.jpg (265KB, 855x1235px) Image search: [Google]

265KB, 855x1235px

>>56180726
>can compete with Intel’s best

>>56180983
AMD promises what they deliver. The fanboys, on the other hand, blow the expectations way out of proportion. It's getting quite ridiculous at this point, you'd think they'd learn.

>>56181181
see >>56181181

>>56181025
You do realize that the zen summit ridge cpu was an engineering sample, right? It probably can't hit high clocks yet, or there may be features that are currently still disabled since they are still in QA.

Even with this processor having such a big handicap, it still edges out intels second best at the same clocks. This shows that it is more efficient even in this unfinished state.

>>56181269
>You do realize that the zen summit ridge cpu was an engineering sample, right?

AMD doesn't have time to redesign and respin before launch. The A0 engineering samples we've seen are final silicon unless they put out a new stepping a few months after launch.

>>56180726
What the fug is with the dusty old Corsair desktops could AMD not afford the new Nvidia cases.

>>56181269
>Even with this processor having such a big handicap, it still edges out intels second best at the same clocks.

AMD has a history of faking demos. The real numbers we've seen from someone with access to the same 8 core SKU was closing to the i7-980 performance-wise.

File: 1459495078743.jpg (4MB, 4912x3264px) Image search: [Google]

4MB, 4912x3264px

>>56181326

>>56180905
underage b&

File: 1b70c8cd-73d1-4cab-bc44-903e46d0351a.png (105KB, 255x345px)

105KB, 255x345px

>>56180726
It takes a god to pull AMD out of the hole it dug

File: 1469448924141.jpg (8KB, 130x193px) Image search: [Google]

8KB, 130x193px

>>56181341
>real numbers
>/g/

Hahaha!

>>56181345
Wow, those SATA ports next to socket (those are the ones interfacing straight from the CPU, rest is from the Promontory chipset).

Man, I would totally want a funky insider-use motherboard like that.

>>56180905

>Source: my dad works at Nintendo

>>56181341

>AMD has a history of faking demos

Citation needed

>>56181025
To make a point about IPC. They could just as well claim that test platform was 3.2 zen against 3.2 broadwell-e.

For all intents and purposes, they could just as well play a video of this shit rendered on Pentium III and speed it up 200 times.

AMD gains 4.8% market share.

'In total discrete graphics, AMD gained 4.8 share points to 34.2% of market by unit volume (based on Mercury Research). In desktop discrete, a subset of total discrete, AMD saw a 7.3 share point increase, rising to 29.9% market share. This is another positive testament AMD’s strategy is working as the company drives forward towards "Vega" offerings for the enthusiast GPU market, which AMD expects to bring to market in 2017 to complement our current generation of "Polaris" products.'

http://www.hardocp.com/news/2016/08/19/amd_gains_48_market_share_in_q1_2016#.V7f6srGxWHs

Nvidiots on suicide watch

>>56181017
they're both incompetent

>>56182068
>thread about cpus
>amdrone shits his gpu pasta

I hope it isnt true.

I just bought a computer, im super poor so i couldnt afford one in the upcoming years...

File: Radeon Technologies Group_Graphics 2016-page-016.jpg (481KB, 3999x2250px) Image search: [Google]

481KB, 3999x2250px

>>56181947

These power consumption numbers can't be real because we now know that poolaris is roughly equivalent perf/watt of maxwell.

>>56182363
>intel core i7 4770k with 4x4GB DDR4
They dun goof'd

>>56182363
those are actually near real performance

my rx480 does not even start its fans at mid settings in most games.

however its also twice the price of a 950, on brand new architecture and mid range not low end.

I honestly would be happy with an AMD CPU/Nvidia GPU rig. Unlike their garbage graphics cards their CPUs were actually pretty neat at the time, they simply fucked it up by being too "future proof" with more cores meme in an industry where future proofing is close to impossible. I really wish them best and hope they will focus on one hardware aspect, preferably CPUs.

>>56180726
im excite

Wanted to upgrade next year anyway for vega

>>56182068
>480RX
>enthusiast
?????????

>>56182704
490

490

retard

>dat 128 threads task manager pic

I wouldn't even know what to do with 16 threads but holy shit i want it.

>>56182704
Wait for the AMD high-end card retard

>he unironically believes that a bunch of pajeets with a couple of million in R&D budget can beats intells 3B+ R&D department and own fabs.

>>56182963
funny ive gone backwards from 8 threads in 2009-2014 to 4 threads since

Kill me :(

>>56182068
Posting update to correct my own posting. Fucking source is unreliable.

HardOCP are a bunch of faggots.

http://www.reddit.com/r/Amd/comments/4yr991/actual_mercury_research_amd_market_share_document/

>>56183048
It's even worse than that, Nvidia spends just for their GPUs what AMD spends on their entire R&D budget

Honest question I would like to know the answer to. Why can't we have single core CPUs that are like 8 gigahertz with 128 megs of cashing?

>>56180726

Zen is not meant to compete with Intel's high end processors, but rather the mainstream market. Stop shilling you fucking Intelfags.

>>56183801
Bigger cache = more time lost when a cache miss occurs.

As for why not a bajillion gigaherz, electron leakage. Shit just doesn't work when electrons intended for one path jumps to another because the material can't handle the heat generated at those frequencies.

>>56183838
>Zen is not meant to compete with Intel's high end processors
what the fuck did I just read

>>56183840
ok. that kinda makes sense. thank you.

>>56183801
>Why can't we have single core CPUs that are like 8 gigahertz with 128 megs of cashing?
That's several questions in one, let's break it down.
>8GHz
A whole lot of reasons but the easiest to understand is that the signal travels at a certain speed and it needs to be able to get to the furthest edges of the chip before the clock cycle is over. The more clock cycles per second (that's the frequency) the less time that signal has to move. That's also why we can't just make huge chips.
>128 megs of cashing
The larger the cache the slower it is, basically. That's grossly oversimplifying it but it's one of the main reasons.

>>56183840
>Shit just doesn't work when electrons intended for one path jumps to another because the material can't handle the heat generated at those frequencies.
I... what? If you're going to give explanations, at least wipe the shit off them after you've pulled them out of your ass.

>>56183889
The smaller the die is, the smaller the insulation area is. The higher the clock frequency, the more heat is generated. The more heat is generated, the less efficient various doped materials are and electrons leak over where they're not supposed to be.

>>56180798
Its valid for me. That means lower proceso. Not everyone aims for the high tier

>>56183867

Sorry man, I'm a little bit too drunk and my banter power is slipping away.

>>56183937
>The higher the clock frequency, the more heat is generated
...

You're jumping several hundred important steps. It's like you've understood the general gist of what's going on but you don't really know for sure. You might have seen it in a picture book.

>>56183986
You want people to write fucking thesis papers on /g/?

It's going to take that to explain what is a quantum phenomenon.

>>56182363
Could you fucking learn to resize?
This place is worse than /ic/.

>>56180937
who is this semen demon?

File: 1470559789682.webm (1MB, 720x1280px) Image search: [Google]

1MB, 720x1280px

>>56184144
checked

>>56181994
>intents and purposes
>implying the saying isnt intensive purposes.

>>56183067
What do you do with it? I have both an i5 2400 and a Xeon e3 1240v2 and there doesn't seem to be much difference desu

>>56184379
Gamez racing sims that sort of thing

>>56184173
I want to bang her mouth

>>56181017
>Nigger it's been 6 years since the last AMD CPU

Uhh AMD released new processors in 2016...and guess what they were shit as usual.

>>56184404
Do you really miss those threads? Personally I only use them for converting video files, other than that I don't notice much difference.

>>56184173
she's just asking to get a photoshopped

File: 79ab0f4b2caf3d35c0135e4c66699f84.png (201KB, 799x574px)

201KB, 799x574px

>>56180905
1. Underage b&
2. You dumb fucking kids will believe anything

>>56184444
Yeah alot of games dont like 4 threads especially CPU intensive games like ACorsa and even normie games like GTA V start to choke

>>56184446
>she

>>56184144

a dude

>>56184519
I thought I read that HT is worse in a few cases but doesn't bring much advantage.

My Xeon is just my home server without gpu anyway so I'll never be able to compare but always wondered how it compares to my desktop i5.

File: habitat67.jpg (83KB, 750x563px)

83KB, 750x563px

>>56181641

And lo, from depths of that hole, did KELLER reach down and wrest into his strong hands the soul of AMD, which he lifted UP into the heaves of computing greatness so that all could enter a state of ZEN.

>>56180905
Are you 12?

>>56184658
>Jim Keller joins Tesla after AMD
>Teslas start exploding

Coincidence?

>>56184173
What? When did Wendall become a hot girl?

>>56184180
>not incense porpoise

Pleb

>>56184732

Tesla couldn't handle the heat. Whereas AMD was smart enough to design their new wraith cooler, because they knew that Keller really is hot shit.

>>56184572
>>56184616
show me a picture of a dude who has eyes that big

Where were you when Intel was kill with Bulldozer?

>>56184767
I really really hope Jim 'Shitwrecker' Keller did his job well once again. I want the CPU market to be shaken up at least a BIT.

Let AMD die.

And then let the legal shit storm of anti trust lawsuits happen.

Intel breaks up.

Two companies inherit Intels know-how.

Competition is restored.

File: a dude with eye that big.jpg (6KB, 161x313px) Image search: [Google]

6KB, 161x313px

>>56184794
here you go

>>56184897
This.

>>56181025
You obviously don't know how engineering samples work.

>Hint, engineering samples are usually a few hundred MHz slower and can't overclock for shit.

>Source I have a lot of ES cpus

>>56184897
>>56185097
Except AMD is actually looking good with their promises on Zen.
However if they fail this time, it's probably over for them.

AMD went from 32nm to 14nm and that's a huge leap in power consumption, heat and performance. I have faith, but only this one last time.

File: tmp_4898-14715779126031252249815.jpg (241KB, 1000x631px)

241KB, 1000x631px

REMINDER
Do not support an Islamic terrorist organisation like AMD

>>56184144
Pistol, she sometimes streams on Twitch. Does some other shit too i don't know.

>>56185186
>she

>>56184415
An 8 core Excavator chip would have kept up with a sub 4ghz 5820k in highly threaded tasks. Not amazing but definitely competitive.

>>56180726
>where were you when AMD caught up to intel from 2 years ago with vaporware that wont drop until intel releases its next die which will outperform it at probably every metric

Sitting right here, I guess.

File: Shut it down.png (659KB, 1041x3977px)

659KB, 1041x3977px

>>56185166
Good goy.

>>56180726
>Hype

Did you all forget what Hype did to bulldozer?

/g/oofs

>>56184069
How bout you get a bigger monitor fag!?

>>56185407
Yeah the bribing is down right scummy, the CPUID stuff is bad, but... You realize this is only the ICC that does this right?

The vast majority of gaymes and programs people use on a day to day basis, are compiled with the MSVC (MicroSoftVisualCompiler). Then LLVM/GCC. Yeah it is shitty what Intel did, but they only did it with their compiler.

The reason bulldozer sucked wasn't Intel, it was that it could only decode 2 bytes per cycle.

>>56185494
The reason bulldozer sucked was because Intel fucked AMD out of a few billion in revenue which fucked their R&D budget.

Also would like to remind you that if Intel hadn't started acting like a monopoly, AMD would have put out something better than bulldozer and Intel would have released CPUs that were better than slightly incremental upgrades these past 8 years.

File: weirdalfoil_2322.jpg (40KB, 350x347px) Image search: [Google]

40KB, 350x347px

>>56185561
>Drones believe this

>>56185494
But look at what AMD has done with ZEN. That was after a birrion dorra payout from Intell.
If AMD had an appropriate market share without the bribery Intell done to have only there chips used, AMD would be way ahead of the game by now with the extra funds pumped into RnD.

>>56180905
Retard will flip out on you but you are right, we are behind ten years in technology compared to the best of today aviable tech for us

>>56180751
Make that 2 years late on 2 fronts, rx480s just a gtx970

File: 1461232096640.jpg (69KB, 480x441px)

69KB, 480x441px

>>56185582
>Zero counter argument
Get a tripcode so I can filter you.

>>56185582
You have to admit though, intel does have a pretty bad monopoly on the market and the it has stagnated a lot because of it.

>>56185591
> AMD would be way ahead of the game

you're implying that only AMD would advance lmao

Whos to say Intel wouldn't? We all know they have been holding back and staggering releases for ages

>>56185614
>because of it.

Intel had already started the monopoly before it all happened, be real.

what Intel did just put AMD even further behind than what they already were

>>56185641
>what Intel did just put AMD even further behind than what they already were

It wasn't always like this, if you were around in the early 2000's. AMD pretty much dominated the cpu and gpu scene back then, it's kinda suspicious for such a successful company to suddenly fall so hard.

>>56185618
>We all know they have been holding back and staggering releases for ages
If that were true, that is only possible due to the Jewish tactics that were (and still are according to your argument) employed.

In an alternate timeline where intell didn't jew the market. AMD would have more $$ for RnD, and Jewtell would have less shekels for such. Given we know AMD is all about delivering the best they can for as cheep as possible, the surge of funds would have propelled AMD's usage in the desktop market significantly. Leading to a dominance

>>56184732
Tesla also has a partnership with Nvidia, im sure they brought a bit of that Fermi magic to DRIVE

>>56185732
>In an alternate timeline where intell didn't jew the market. AMD would have more $$ for RnD, and Jewtell would have less shekels for such. Given we know AMD is all about delivering the best they can for as cheep as possible, the surge of funds would have propelled AMD's usage in the desktop market significantly. Leading to a dominance


This is pretty far fetched man, even for a drone.

>>56185775
Not really. Intell had to bribe PC manufactures to NOT use AMD chips, in order to get the advantage they have now.

>>56185618
Intel is doing that because it can because they fucked AMD illegally, that's the exact argument that we are making.

who cares if intel bribed, cheated etc? they have the money to do that, it means they're superior. end of story is intel/nvidia are ALWAYS at the top of the charts. who cares how or why. just buy the best products.

>>56185831
>Bribe

it's called undercutting the competition or predatory pricing, offering prices too cheap to refuse.


Predatory pricing is illegal under anti-trust laws.


If Nvidia were to release GPUs cheaper than AMD in the same performance bracket they would get hit by lawsuits too because businesses with substantial market shares are more vulnerable to antitrust claims.

>>56185996
see image attached to
>>56185407
And
http://www.extremetech.com/computing/184323-intel-stuck-with-1-45-billion-fine-in-europe-for-unfair-and-damaging-practices-against-amd

>The EU found

>That Intel paid rebates to manufacturers on the condition that they would buy all (Dell) or nearly all of their CPUs from Intel.
>That it paid retail stores rebates to only stock x86 parts.
>That it paid computer manufacturers to halt or delay the launch of AMD hardware, including Dell, Acer, Lenovo, and NEC.
>That it restricted sales of AMD CPUs based on business segment and market. OEMs were given permission to sell higher percentages of AMD desktop chips, but were required to buy up to 95% of business processors from Intel. At least one manufacturer was forbidden to sell AMD notebook chips at all.

Intel was busted and fined 1.4 billion dollars for BRIBING PC manufacturers.

This happened and was proven in the court of law.

>>56185948
>they have the money to do that, it means they're superior

Oh, how about the many other markets where only one company has a monopoly on the entire market and thus there is zero progress? How about Comcast and Time warner practically owning the entire US when it comes to ISPs, which is why America's internet has been so bad for years? A monopoly is never a good idea, regardless of who is it for.

>>56186070

This sounds like the normal anti-business drivel out of the EU. Are you really trying to side with an organization that thought importing ~1 million syrian terrorists was a rational idea?

File: 1286792797780.jpg (12KB, 279x267px)

12KB, 279x267px

>>56186113
>throwing the baby out with the bath water.
How shilled are you?

>>56186113
Source? There's literally zero source out there about AMD trying to bring in refugees or anything from Syria. Nice try Netanyahu.

>>56184642
HT results in the processor switching to a different thread on a pipeline stall from a cache miss or branch prediction miss, if you have some shitty python program with cache misses left and right thia could easily double the overall performance. If you dont have any of those then you have the problem that the cache is shared for both threads per core resulting in maybe 5% less performance.

>>56181345
What cases are those? I don'e even like memesair cases but those are cutes.

I love AMD (In GPUs is the only I have had) but I can't defend them right now since they're not giving us shit.

Untill they break the "we will do it guys, for sure" AMD is kinda dead in market.

If they release something, and even more, if they release something that Intel or Nvidia won't surpass it to the next week, they won't be taken serious.

But, this is my opinion, of course, not some sort of truth.

File: 1471656954809.png (415KB, 646x438px)

415KB, 646x438px

>>56186460
>If they release something, and even more, if they release something that Intel or Nvidia won't surpass it to the next week, they won't be taken serious.

You still have to consider the price performance point.
I cant afford to pay $200-300 more for X performance over AMD when it's something that only really applys to benchmarks. Sure 10-20 fps more/calculating pi to the 10th %15 faster is great. But for a consumer grade market, you can't tell the difference for a day to day use.

All the commie bullshit in the world couldn't fix the separation between Intel and AMD. And even if they got punished, let's not forget that was years ago. And even lacking the R&D funds, the Bulldozer arch wasn't very good. It's not horrible, but AMD can do way better.

And let's not play like AMD is a mega underdog struggling under the oppressive thumb of the Intel monopoly. They make the internals for video game consoles. Which is hella cash there dawg.

They have their thumbs in a lot of pies. But they're not drowning. Hopefully Zen can bring AMD back on the CPU radar for enthusiasts.

If we really wanna punish Intel, why not force them to give up their patents and trademarks on their CPU designs to the public domain? If they're so good, let them start from scratch with a whole new slew of competitors ready to snatch up those designs.

Fuck everybody, including myself.

>>56187270
>They make the internals for video game consoles. Which is hella cash there dawg.
It's not really that much, enough to keep them afloat at best.

>>56184767
When will stock coolers stop being so compact?
I get that they do need to fit in all builds, but could we at least have a choice of a proper tower cooler? I'm more than happy to pay $20 more for a good back up when my water pump dies.

But not to stomp on stock fin and pipe coolers, they work fine. But the slabs of aluminum with a fan attached need to go.

>>56187309
I actually think that the new AMD cooler is pretty good as far as stock coolers go. It is close to performing at 212 levels.

Now, if we could get OEM chips *without* a heatsync for $10-15 less, I'd be all for that, but if they're going to bundle them regardless, AMD is at least moving in the right direction.

>>56187309
>I'm more than happy to pay $20 more for a good back up when my water pump dies.
I mean, at that point you could buy an aftermarket Hyper 212 Evo or equivalent.
The new AMD stock cooler is decent though.

Why does AMD have to compete with both Intel and Nvidia?
Can't they just focus on one thing?

>>56187391

They have to compete with Intel because their bread-and-butter is CPU design. They compete with Nvidia because they bought Radeon for $5.4 billion, and to do nothing with it would be a huge waste.

>>56187326
>*without* a heatsync
>heatsync
>sync
K...
> $10-15 less
I think without a HEATSINK it would cost even less. Because of the heavily reduced shipping costs.
You could fit 15-20? cpu's in a fx-8350 box without the cooler.
So you may be looking at $25-$30 less when getting it from a big retailer.

I'm gonna be using the fan out of my 8350 for a south bridge cooler on the 970.
So I'm not really fazed personally by the extra costs of the added hardware. I just see it as backup and useful parts.

But for those that use air cooling, I can see why you'd want to be able to only buy the chip.

>I actually think that the new AMD cooler is pretty good as far as stock coolers go
Same I was really surprised by how quiet that fan was from the 8350 (non wraith?)
Even tho all the newegg reviews were uppidy about it.
I can't even hear the thing when it's in the case. Fucking old HDD's make more noise.

So they down clocked the Intel chip (because the AMD chip is maxxed out and cant be overclocked)

Sorry, but AMD has to deliver a sub $150 throwaway cpu for me to even consider going over to the poorfag side. Let me know with AMD can hit 5ghz 24/7 without the need of LN2 and fire insurance.

remember how AMD said that Faildozer was going to destroy only get buttraped by Nehalem?

Yeah i dont think so.

Poorfag products; poorfag results.

>>56187340
>I mean, at that point you could buy an aftermarket Hyper 212 Evo or equivalent.
True, but then I'd have to go and buy one.
Where as if I just buy the cpu and it comes with one. I'm saving on time and shipping.

>>56187435
1) I cant spel wurth shet.
2) I'm assuming that consumers wouldn't reap the full cost reduction associated with not including a CPU cooler, which is why I said $10-15. I mean, you can get a Coolermaster Hyper T2 for $17 from Amazon, and that's roughly equivalent in terms of material and weight to the new Wraith stock cooler. Getting a CPU minus cooler for $25-30 under the cost of the CPU with cooler is optimistic, I think.

>>56183889
>>56183986
>I'm going to point out everything you said and claim it's wrong, but I'll offer no information to correct it because I'm a shitposting wannabe

>>56185698
True, until intel said "hey rather than just making home computing CPUs why dont we get in the gaming market?"

Intel was bigger than AMD even at that time. Intel just decided to pump a fuckton of money into and came out with CPUs the buttfucked AMD six ways till sunday, and the rest is history

>>56187458
*destroy sandybridge

>>56182363
>GTX 950
And which Polaris are they comparing to it?

>>56187519

That's not what Intel said. Intel said "Fuck, AMD has a better product that it's selling for below our own production costs. They could kill us in the long run. Shit, what now? I know, let's sell our own products to OEMS at below our own cost, under the proviso that they can't buy anything from AMD! Since we have more cash, we'll just wait until AMD is out of the market since they won't have OEM customers."

>>56187462
True, but I mean, I have a Fry's electronics that's like 10 minutes away.
Bought a Hyper 212 Evo a year ago there for like $25.

>>56187477
>consumers wouldn't reap the full cost reduction associated with not including a CPU cooler, which is why I said $10-15
That's fair enough.
Also I'd like to say maybe keeping a stock cooler packaged is also good for open markets.
like I mentioned that reduced shipping savings would only be applicable for big buyers. So your local computer shop wouldn't be able to compete.

>you can get a Coolermaster Hyper T2 for $17 from Amazon,
That's pretty cheep. But I don't think it would stay that way for too long is you were able to solely purchase the cpu. I'd think some price hikes would happen with the upped demand.

But that's semantics for now.

File: single threaded.png (29KB, 731x390px) Image search: [Google]

29KB, 731x390px

Some of the stupidity in this thread must be either shilling or pretending.

>>56180798
>It can compete with Intel's medium tier CPUs at best.
Are you fucking retarded? Intel has 1(one) core that they gimp(and relatively at that) per generation. If you actually compare the single thread performance of say, all Skylake processors, you'll notice that they're pretty close. Pic related.

>>56181025
Because it's a fucking engineering sample.

>>56187458
>So they down clocked the Intel chip (because the AMD chip is maxxed out and cant be overclocked)
Kill yourself.

>Let me know with AMD can hit 5ghz 24/7 without the need of LN2 and fire insurance.
Can you point me to any CPU that can hit 5 GHz at sane voltages? Just any CPU at all. Also, >falling for the clock speed meme

>>56186460
> if they release something that Intel or Nvidia won't surpass it to the next week, they won't be taken serious.

yeah okay. meanwhile their stock has jumped something like 300% this year alone.

>>56184816
Where Keller did most of his work was conceptualizing the design stages and managing the team.
He set up a foundation for AMD to advance but we have no idea how well they're going to do it.
Zen+ rumors of 15% better than Zen puts it near/above Skylake octocores if that Blender test is anything to go by
Moving on to Kaby Lake,
5%? More features most people dont need? Intel's 4th sidegrade in the past 6-7 years? Zen+ competing with cannonlake? No thanks.

>>56185234
No. The 5820k is a 6 core processor. Excavator cores are much smaller and there is only 8 of them in your "what if". That doesn't overcome the deficit in threads nor are you throwing more hardware at the problem. (5820k has 12 threads vs 8 on the excavator). Thread to thread, AMD could win, but at such a huge deficit it's not going to happen.

If it was competitive, it would in fact be amazing. But it's not, so it's not.

>>56187613
>300% increase
Thats easy to do when your stock performance has been dog shit for decades.

that would be like "hey our car company is so shit we sell 3 cars a year for the last 20 years, but this year we sold 10!!, look at are massive and totally impressive comeback!!!!"


As an investor i dont consider it impressive. Now a 40% increase when your stock has been doing exceptionally well for decades is impressive.

File: 1438039861843.jpg (5KB, 364x138px)

5KB, 364x138px

>>56187605

>>56187537

Fucking this anon. I don't think that most /g/tards realize how shitty Intel behaved during the late 90s/early 00s.

>>56187790
I mean AMD would've done the same thing if they had thought of it first.
They're a company and their #1 priority is to make products that bring in revenue.

>>56187711
Zen+ is absolute pure speculation. IPC on Kaby Lake is probably almost identical to Skylake but clock speeds have increased significantly. In any case we have only a vague idea of Zen IPC from basically two benches, and no idea how high it'll clock, probably a whole lot less even if it's significantly faster than the engineering samples we've seen so far.

>>56187776
>As an investor I don't consider a 300% increase in 8 months impressive

You sound genuinely retarded.

File: 1469883273122.jpg (73KB, 960x960px)

73KB, 960x960px

>>56187808
>I mean AMD would've done the same thing if they had thought of it first.
Based on what?
What have AMD done or tried to do that lead you to this conclusion?

>>56187808
Intel broke the law, anon. AMD did not. If Intel hadn't illegally colluded with OEMs, AMD would never have had to spin off Global Foundries, and could be competing with Intel fabs instead of needing to farm it out.

>>56187812
Well the Zen design team said "Zen+ on track for two year cycle and +15%" when they presented marketing claims.
If AMD doesn't get together a team to plan for after Zen they are in the same trouble Bulldozer started.

>>56187833
>Intel broke the law, anon. AMD did not.
And look who is ahead now.
It's dirty practice, and Intel definitely shouldn't have been assholes like that.

But any company that wants to make it big has to take risks such as that and in the same position AMD probably would've done that.

>>56187829
It's really sad that I can't even tell if you are being sarcastic based on the image you used.

>>56187865

You do realize that Intel's breaking the law had the effect of screwing over consumers, who were forced to buy lesser products for greater amounts of money, right? And it eventually led to AMD falling behind to the point where Intel became a defacto monopoly in high-end CPUs, resulting in them being able to charge a very high premium for them, right?

Intel's illegal actions fucked over you, anon. And you're defending them because 'if they wanted to make it big, they had to take risks?' Crazy.

A real risk was when Bob Noyce broke off from Shockley and then Fairchild to found Intel and took with him all the top engineers he was working with. That was good for them, good for the industry, and good for consumers. Intel colluding with OEMs to eliminate competition in the marketplace? Not so good for anyone except Intel.

File: 1470722584895.jpg (63KB, 652x484px) Image search: [Google]

63KB, 652x484px

>>56187865
>My Father is my Mothers rapist.
>He took a illegal actions in order to have me.
>But that's fine I only inherited his down syndrome.

>>56187904
>not fighting intel in court when it started
>Get buttfucked and just takes it
>then AMD goes full blown retarded and buys ATI instead of investing in their cpus.

AMD deserved what it got.

They've had enough time and money to really make something in the last ten years to blow intel out of the water performance wise and have NEVER delivered

>>56188016

They did fight it in court, you idiot. Courts take time (years and years and years when you include appeals) to come to a decision. AMD won its court cases, both in the EU and the USA. By then it was too late though.

Do you think that innocent people who get murdered by gangbangers deserve what they get too?

>>56188016
>>not fighting intel in court when it started
http://ec.europa.eu/competition/sectors/ICT/intel_provisional_decision.pdf

1. Commission procedure
(5) On 18 October 2000, AMD submitted a formal complaint to the Commission under
Article 3 of Council Regulation (EC) No 17/62, First Regulation implementing
Articles 81 and 82 of the Treaty.8
(6) On 26 November 2003, AMD submitted a supplementary complaint under Article
3 of Regulation (EC) No 1/20039
providing new facts and making new allegations.
(7) In May 2004, the Commission launched a round of investigations focusing on
allegations contained in the supplementary complaint. Within the framework of
that investigation, in July 2005, the Commission, assisted by several National
Competition Authorities, carried out on-the-spot inspections under Article 20(4) of
Regulation (EC) No 1/2003 at four Intel locations in [...] ([...]10 [...]), [...], as well
as the locations of several Intel customers [...].

AMD has fought for many years. Whilst still producing a product worth using. EVEN when there budget was slashed because of intells JEW tactics.

>Cuck me harder daddy

>>56188054
don't bother, guy's either trolling or a moron.

File: heebtel2.png (305KB, 929x1320px) Image search: [Google]

305KB, 929x1320px

>>56188054
illegal stuff is fine, if it is the eternal merchant doing it
remember the six million

>>56188092
>Whilst still producing a product worth using
>products got buttraped by cpus 2 generations older

Sure

They still mismanaged themselves into oblivion and still went full blown retarded by buying ATI instead of focusing on their CPUs


They deserved everything they got.

>>56188094
Sadly we have to. People come here for information. Only to receive shillary.
I wish this place didn't have /v/. We wouldn't have to put up with this bullshit.

>>56184794
>>56185042

Fuckings Kek.
1010post

File: 1471647245885.jpg (57KB, 728x546px) Image search: [Google]

57KB, 728x546px

>>56188118
see
>>56186681

>>56188118
their gpus and entry apus are literally their only products worth buying until zen comes out. but what do you expect from an idiot defending illegal market manipulation.

File: 1416392641096.jpg (53KB, 700x836px) Image search: [Google]

53KB, 700x836px

>>56188150
>price to performance ration
>not wanting the best of the best

Sorry, but i pay for products that come in first place.

You should sit down with your wife's son and explain to him why striving for 1st is more important. Otherwise, he'll end up being a cuck like you.

File: good goyim.jpg (11KB, 208x242px) Image search: [Google]

11KB, 208x242px

>>56188201
>Talks about others being cucked
>Is literally being cucked at this very moment by lemerchantface.intel.jpg

>>56188201
>you see son what you do is add a whole bunch of unnecessary tessellation inside objects and under the map. That way you receive kickbacks from nVidia

>>56187789
The shills are strong nowadays, Intel must be scrambling for something like this
Still pretty much within the expected results that Su promised, maybe a little bit on the better results side

>>56186681
nigger are you telling me that I, as someone who's had a pc for almost a year, would not be able to tell the difference in performance in my gaymen whether they are driven by an rx480 or a fucking $1200 TitanX? I agree that some people can't/wont pay more for more performane, but you must be the least perceptive person to not notice the difference between different tiers of GPUs. I mean its either same setting and more frames, higher settings and same frames, or higher settings and more frames, not rocket science

>>56188211
>settles for second place
>hurrr im totally not a cuck.

>>56188267
>changes the settings in some gay gaym
>hurr this totally negates the fact that i'm raising another man's son

>>56180905

Was your chest puffed up as you typed that? Did you even get a little bit hard? Do you realize that creaming in your pants for a company/product is the lowest of lows? Stop basing your identity and worth based on what brand you choose to back. You're not associated with it other than being their sheep.

i'm pretty excited for this tbch
might make AMD viable again, mostly because you can still get away with an overclocked i5-3570k for most everything out so far

File: stock-hsf.jpg (69KB, 600x514px)

69KB, 600x514px

>>56187309
The Phenom II's had nice stock coolers
the fan was loud as fuck though

>>56180726
This info is meaningless without the fucking price.

>>56188324
No kiddin', they'd hit 5k rpm over fucking nothing too.

>>56185147
>Except AMD is actually looking good with their promises on Zen.
Sure. Two years too late.

Again, doesn't matter if Zen delivers 100% of the performance promised if it's so fucking late that Intel's already a generation (or two) ahead.

>>56184415
>Uhh AMD released new processors in 2016...and guess what they were shit as usual.
All of them are based on an arch revision that's 3+ years old, and all of them are mainstream APU's
>>56182363
That's not even a demo, it's a slide
The figures are for Polaris 11, which isn't even released on any product yet
Polaris 11 will be a much lower clocked part than Polaris 10, with much lower voltages and thus much lower power consumption
>>56183048
The CPU division doesn't even have pajeets nowadays, the pajeet team that designed Bobcat was poached by Samsung years ago
Also, AMD has always been an underdog to Intel's budget, and has even beaten them despite the smaller budget
>>56185147
>However if they fail this time, it's probably over for them.
AMD's IP portfolio is big enough for them to live as a licensing company
Just see Cyrix and VIA, they're still around
>>56185494
>The vast majority of gaymes and programs people use on a day to day basis, are compiled with the MSVC
All the performance stuff back then was compiled with ICC, this was demonstrated back then with benchmarks that showed the gains AMD had when spoofing their CPUID
Nowadays MSVC is fast enough and much cheaper than ICC ever has been, and GCC can beat ICC sometimes
Only GNU haters and Apple use LLVM/Clang
>The reason bulldozer sucked wasn't Intel, it was that it could only decode 2 bytes per cycle.
Don't you mean two instructions per cycle? Do you have any idea about what you're talking about?
The reason the first Bulldozer sucked so much was that the process didn't deliver, being leaky as fuck and seriously limiting the high clocks Bulldozer was designed for
>>56185561
>The reason bulldozer sucked was because Intel fucked AMD out of a few billion in revenue which fucked their R&D budget.
That's the reason the first Phenom sucked
>>56184732
After Keller joined Tesla in their Autopilot division Tesla had the balls to release Autopilot cars
Keller is out of this world

>>56186113
>>56186070
>>56186141
>>56186164
EU is pretty good at being fucking retarded with shit like this. They're trying to fine Google for anti-competitive practices for their play store as well, since someone like Amazon doesn't have their shitty bloatware version of the playstore installed by default.

YET THEY'VE DONE NOTHING TO APPLE, whose mobile OS cannot even support a competing app store.

>>56188324
>>56188338
So what you mean is... their stock cooler was shit.

>>56188264
Well duh. I'm just calling you a fuckwit for sponsoring a high priced manufacturer at this point in time.

I'd rather spend my money with AMD to boost there RnD right now, because most games look shit on either chipset. Regardless of frame rate. (yes games look better now, but still not where I'd like it to be to pay a fuckload on something that'll only last a year or two)
By doing this I'm hedging my bets on future games that'll require greater computing power.
AMD has never fuck me over on price point. So if I can have that trend to continue and get a product as good I'll buy AMD.

Sure if you don't give a fuck about prices spiraling out of control buy nVidia. You're just helping me hedge by spending your shekels with them.
But I'm not like that, I don't want the AAA PC gaming market to be solely for those earning 80k a year.

>>56185618
>We all know they have been holding back and staggering releases for ages
Not really, they tried for years to get into mobile and failed to the point they are going to license ARM cores now
>>56187776
>that would be like "hey our car company is so shit we sell 3 cars a year for the last 20 years, but this year we sold 10!!, look at are massive and totally impressive comeback!!!!"
Yeah, this is totally what happened and the reason people are buying AMD shares, just a minimal surge in prospects made investors buy it, can you believe how genuinely retarded are they?

>>56188358
Another expert heard from. A 3-10% difference between Skylake and Zen amounts to shit if Intel's still selling i3's at a price that AMD is selling their i7 equivalent, let alone their broadwell-E shit that seems essentially comparable at a thousand fucking dollars.

>>56188420
>A 3-10% difference between Skylake and Zen amounts to shit if Intel's still selling i3's at a price that AMD is selling their i7 equivalent
Well, are they?

I have a feeling we won't be so lucky

>>56188456
The 4/8 Zen skew seems to be the absolute lowest so there's a decent chance. I'd be shocked if it was more than an i5 at most.

>>56188358
>Two Years too late
The CPU market is extremely stagnant.
2 years is a few watts off the TDP and a few hundred MHz higher clock speeds.

Intel has been sitting on their asses, AMD finally being able to catch up to their new 6000 series Broadwell-E processors means a shit ton.
32nm to 14nm was a huge leap for AMD.

if amd will not have a component embedded similar to intel's management engine, they'll get my money

>>56188404
you sure are mad. idk why you think I'm an nvidia fanboy, but I guess amd fanboys are like that. at the time I made my pc the 980ti was the best card to do what I wanted. right now, it seems like amd's hardware level async is awesome in dx12 games. if nvidia can't keep up when they release their 1100 series I will strongly consider jumping ship even though I've got a gsync display

>>56188512
>if amd will not have a component embedded similar to intel's management engine, they'll get my money
Not a chance, see the libreboot and FSF comments on those things
Anything past 2013 by AMD has such anti-features, there's no going back from such ride

surprise me AMD

I want to be able to buy a KabyLake cpu for less, thanks to AMD's "competition"

>>56188659
Eh, I wouldn't even recommend AMD Just yet for dx12 just yet.
>DX12 has no game
So I'll have to say wait until the next release of cards mid-late next year, with the new memory set.
Right now it's not really worth going for a top tier card. It's a bit late in the gen.

File: 1470958051315.jpg (897KB, 1080x2609px)

897KB, 1080x2609px

>>56188755
It's much faster than dx11 adoption.

>>56188769
I'm guessing that'd be due to a win 10 rep.
But still, that's not really much, unless your card is past its prime already. and are desperate to upgrade.
I've only an R7 370 myself, and thought about the rx480. But with ZEN and the new memory set next year. I'm saving my dosh to go all out.

>>56188755
that's why I'm waiting. right now it seems like async works really well with dx12, which amd does well and nvidia doesn't. despite what nvidia says, they have yet to implement async at a driver level, and currently, neither maxwell nor pascal can handle it at a hardware level. I'm not sure how nvidia will fix it, but it seems like they'll have to make considerable hardware changes. they should be able to do that if they invest billions in whatever comes after pascal, but who knows.

I know that battlefield 1 will have dx12 support and some games like doom are patching in support for apis with async support. I'm going to wait till the 1100 series of gpus to upgrade for sure (it'll allow me to get a catback exhaust for my meme car BRZ), and I'm excited for what vega will be bringing to the table

Any one know the ball park prices on ZEN and Vega?
Even the AM4+ socket?

>>56188201
>implying anyone on 4chan has ever been the best at anything
Son, if you're here, you're a cuck or a wizard. Winners don't use 4chan.

>>56188991
Top end Summit Ridge will probably be $500
The big Vega GPU will launch with the same MSRP as the Fury X

>>56189154
I'm guessing that'll be the 32 core server type?

Done some scrounging
http://www.pcadvisor.co.uk/new-product/pc-components/amd-zen-processor-release-date-price-specs-features-3643552/
FX-Zen Quad Core: £80 incl VAT
FX-Zen Six Core: £120 incl VAT
FX-Zen Eight Core: £180 incl VAT ($300AUD)
APU-Zen Quad Core: £100 incl VAT

These seem a bit under priced. But hey AMD.
If it live up to close of what's expected, I'll be shelling out for the 8 core no worries.

>>56189233

no way will this be real. Amd will try to compete but keep prices as high as possible

>>56189233
>wot
No, genius. Socket AM4 only supports 8 core Summit Ridge. Consumers are not getting enterprise Opterons on a consumer socket.

How can anyone be so stupid?

>>56189253
I dunno dude. If I were AMD I'd price it as cheep as possible just to ass rape Intel back while keeping it legal.

>>56189233
the 32 core will be at least $2000

>>56189335
Closer to $5,000.

>>56189346
"at least"

>>56189346
Nah, probably $7,500.

>>56188506
>32nm to 14nm was a huge leap for AMD.
I'm going to be so mad if they're pigeonholed into using GloFo because of the wafer requirements but I suspect they shit out polaris just to eat up those requirements as quickly as possible.

>>56189393
There is absolutely nothing wrong with Samsung's 14nm LPP being ran at GloFo Fab8.

>>56180726
"while the other featured a new Intel Broadwell-E Core i7-6900K CPU. While the Core i7-6900K can run up to 4GHz on some workloads, AMD conducted the test with both CPUs locked at 3GHz."

So its faster clock for cock but still 1/4th slower.

>>56189512
still got like half a year till they're released, they have some time to get the clockspeeds higher.

>>56189512
>>56189548
They ran each chip locked at 3ghz just for a clock normalized comparison.
3ghz is not in anyway a limit for Summit Ridge.

>>56189355
>>56189351
>>56189346
>>56189335

highly doubt it. I have a pretty new 32 thread server and its whole retail price is around $2000-3000, not including disks.

I'm guessing their server cpus would be at most $1500 a piece if they want to compete.

>>56189427
if it doesn't hit the same clocks as intel it'll be shit on anyways.

>>56189607
Some 16c Piledriver Opterons sold for $3,000~
A 4 die MCM with 64t is not going to sell for less than half the price of an intel chip with a fraction of the throughput. 32c/64t Opterons will sell around $5,000, maybe even higher.

Intel's 165w 24c/48t Xeon E7 has an MSRP of $7,174.
Thats what 32c/64t Opterons are competing against.


>>56189611
Its not even remotely that simple. Real clock speeds haven't been what you see on the side of a box for the better part of this decade. They're running through a dozen different pstates at any given time unless you override them overclocking.
Intel's latest turbo strategy uses sensing paths to find the core with the least transistor variation to determine which can clock the highest the longest in serial workloads. The i7 6950X can do this up to 4ghz on that one favored core for short bursts.
Summit Ridge can get off an entire CCX to push clocks higher on 4 cores at a time, or gate off an entire CCX and 3 cores in the remaining CCX to let a single thread use all of its available TDP. The exact clocks its turbo achieve and its given behavior are yet to be seen.
Its not as simple as listed base clock and max turbo.

>>56189512
Intel's turbo boost has been around for so long I can't believe some people don't understand how it works.
Chips like the 6900K only operate at 3.8-4Ghz during heavy single core single thread work.
The all core turbo under heavy loads like rendering probably drops the processor down to 3.4-3.5Ghz

Zen engi sample has a boost of 3.2Ghz.

Skylake-E can typically overclock to 4-4.3Ghz on all cores, I wouldn't doubt Zen can possibly hit 3.8-4Ghz.

With speculative Haswell/Broadwell performance, 8 cores and not so far off clockspeed, 60-80% the price would be a likely move. I wouldn't be surprised to see bins rated for 3.5+Ghz and 125-140w TDP if the 3-3.2Ghz Zen is their 95w part. Wishful thinking on my part.

good, maybe intel skylake-x will be cheaper. Still noit buying AMD:

>>56189762
> getting a memelake
> ever

and i have a memelake cpu on my laptop. Shittiest one I've ever had.

>>56184173
fucking whores mang. she wants it

>>56189694
The most thermal budget any single stock speeds core at this litho stage is ever going to use is about 10w, are the power densities really so high you can't run two cores at max?

I still need to see how the L3 behaves before making claims and passing judgement

>>56189823
Roughly 10w is what Broadwell and Skylake use per core when fully loaded in a normal workload, not counting uncore and other things.
Broadwell cores are roughly 7mm2 each, I assume Skylake cores aren't much different.

If they were allowed to pull 20w each we'd be looking at an extreme amount of heat generated in a very tiny area.

>>56189781
I've got devils canyon just so satan appears in my bios. also skylake wasn't around when I buildt my pc

>>56185561
they tried something new and risky and it didn't pay off

>>56180905
>My dad works at Nintendo!

>>56181341
I think you're confused.

https://www.youtube.com/watch?v=TAGrPyMKA_k

This isn't the first time they've done this, and won't be the last.

>>56190361
>"real DX11"
>VLC fullscreen controls pops up

The actual WAT.

And was the guy doing the demo ACTUALLY a jew?

File: hd4000 failures.png (3MB, 1920x1080px)

3MB, 1920x1080px

>>56190397
They faked DX11 performance on their IGPs. They actually did it a couple times and got caught each time.
Yes, hes a Jew.

>>56186070
You know intel still hasn't paid that fine to AMD, right? They're tying it up with appeals, they're currently on their 4th appeal.

I think they aren't even adjusting for inflation, intel are such fucking jews.

>>56186397
Obsidian 750D airflow edition
$180 meme case

File: ZEN2.jpg (96KB, 1200x696px)

96KB, 1200x696px

Reminder: Work had already begun on the Zen+ core when AMD gave their 2015 FAD presentation.
Zen+ based parts will be either 14nm SOI FinFET, or 7nm SOI FinFET.

>>56190614
I just really hope Zen overclocks well. I enjoy overclocking.

>>56191038

One thing that i'd like to know (which basically nobody will until it launches) is IHS or TIM?

>>56191072
Shit, I'd say it'll be TIM. Plus it gives us more room for error when replacing it. Hence more sales. But I haven't really known of issues with AMD doing it. it only seems like an intell issue with using too much silicon when fitting the sink making it cock up one corner causing core 6? to gain 10-15C

IHS would be nicer, but that may cause more problems in manufacturing.

>>56191038
8 core parts should be able to hit 4ghz, APUs will probably clock a little higher, but they likely won't scale much beyond that.
14nm LPP provides an uplift in average fmax over 28nm HPP.
The chips to look at for reference are Carrizo/Bristol Ridge and Kaveri/Godavari.
Kaveri is 28nm SHP, a custom tuned process in 28nm bulk to favor frequency.
Carrizo is cheaper generic 28nm HPP, but AVFS helps it regain some lost clocking potential.

Average Kaveri OC is about 4.4 to 4.5ghz, chips hitting higher than that are unicorns. Stock clocks however hit 3.7 to 3.9ghz which isn't terrible.
Carrizo isn't available in an overclockable desktop package yet, but AMD kindly showed 1:1 accurate data on its clocking behavior at ISSCC. On average it outpaces 28nm SHP Kaveri in efficiency right past 3.5ghz. We don't know its average OC limit, but with voltage scaling losing to SHP at higher frequencies I'd expect them to have a clock wall a bit worse than Kaveri.
From that we know if Carrizo were implemented in 14nm LPP silicon it would clock higher than that in its peak perf/watt range.

Obviously Summit Ridge is an entirely different core so it can't be used to make a direct comparison, but we know that engineering samples can hit 3.2ghz without issue. I suspect Zen at 3.5ghz would be pretty easy to hit stock for a decent binning. If voltage and clocks scale similarly to what we see in Polaris then power consumption will increase very linearly, then the chip will just hit a wall and not function at higher frequencies. Power and heat aren't limiting factors, the transistors themselves are.
So I'd place my bet on 8c Summit Ridge chips hitting 4ghz, not much beyond. Maybe 4.2ghz~.

>>56191072
Doesn't matter much either way if done properly.

>>56191198
but will APUs have L3 cache this time.

>>56191219
Its likely.
I think the Summit Ridge die is around 125mm2-150mm2.
Raven Ridge's IGP is apparently 12CU, that should fit in the same relative die area as one Zen CCX, so the APUs would have approximately the same die size as the CPUs.
I don't think they'd feature a Zen CCX without L3 cache, seems intrinsic to the design.

File: 1469420730601.gif (495KB, 600x600px) Image search: [Google]

495KB, 600x600px

>>56191198
>>56191272
Cool. Thanks for the insight Anon.

Guess I may even look into some form of chilled cooling to get a possible edge. But I'll have to wait until I see what the board layouts will be in order to make precautions for condensation.

But even then I might only get 0.120-0.150ghz With how finite thing are getting. maybe the .2 bracket, if I get sub zero. (Conservative numbers I haven't into chilled)

>>56185582
Intel lost an antitrust lawsuit over it.
They still owe over a billion dollars in settlements.

>>56191712
I had a look into that. apparently that goes to the EU to make sure that retailers stock AMD laptops.
I don't know how that'd be implemented, where be it subcitys for buying AMD at the retailer level, Or some sort of inspector.

But there was another suit in the US that intell had to pay AMD. Can't be fucked looking into that right now.

>>56191807
>stock AMD laptops
This is anti-consumer. AMD always had shit mobile chips.

File: AMDvsIntelretailPC.gif (30KB, 896x553px) Image search: [Google]

30KB, 896x553px

>>56191712
Pic related
AMD was going to surge past 50% consumer X86 market share. Intel's OEM bribery, faked benchmarks, biased compilers, and outright fraudulent ad campaigns crashed AMD's profits.
If the world were just intel would be paying AMD well over $10,000,000,000 in damages. Their market growth immediately died, and revenue more or less began to stagnate.

They've had to trim the fat constantly, and have barely been able to maintain positive cash flow by lowering overhead. They've sacrificed R&D and man power, sold buildings, all to keep the core business running. All because of intel's kikery. They owe AMD 11 figures.

>>56191829
I can't say for sure hence "apparently". Because of what was outlayed in the report they also jewd the desktop market heavily.

But even if it's subsitys on AMD's laptops, it'll be the EU that gets it. So it'll be a big win for their lower class country's that really vouldn't aford intel to begin with.
But even still, it's all greater revenue for AMD that'll hit R&D. so it's win win win for westerners, the EU and AMD.

>>56188373
EU is trying to make money here, Apple probably doesn't have much to pay.

>>56184144
A lovely man. desu

>>56191989
more like apple would just tell them to get fucked because they sell to the much bigger asian market.

File: Zen team.jpg (75KB, 715x573px)

75KB, 715x573px

Zen is AMD's P6.
These are the people responsible for it.

>>56192383
POOJEET

>>56190460
Wait, is this true!?

>>56192650
Yes. They'll appeal to the EU's highest court next.

Just ordered a 6500+DDR4 RAM and a good motherboard. Fuck waiting until February.

so many insecure brand faggots itt. get a fookin life m8s and stop worrying about what other people put in their computer

>>56181025
Only an idiot would show the competition the best clock he can achieve 5 months before release, especially considering Intel will also launch a product line near in the same period.
They chose 3ghz so that no one could guess what their final frequencies will be.
That said, rumors say boost clock will go near 4ghz.
i7-6900K has a maximum boost clock of 3.7ghz and a tdp of 140w.

>>56192885

My name is The Stilt and i'm an armchair engineer and thus clearly know that Zen is shit because I made up a bunch of numbers while shitfaced on blind ivan's vodka one night and thus can prove Zen can't clock past 3ghz.

File: GCN vcore-clock.png (37KB, 1171x703px)

37KB, 1171x703px

>>56193025
Hes incredibly stupid for a smart guy.
Before Polaris launched he was saying GCN needed to clock to 1600mhz stock or AMD couldn't compete. Since then hes been on a crusade against Samsung's 14nm LPP, desperate to find some fault with it, outright ignoring empirical fact.
I guess hes got an issue with admitting when hes wrong. Typical of forumites.

He himself even compiled a charge of average voltage and clock curves for various GCN cards, showing an increase in fmax over 28nm HPP, and he himself tries to use it to say AMD changed clocks last minute on the RX 480. Ignoring that a 32 ROP design with low clocks never would perform anywhere near the 64 ROP Hawaii based R9 290X, and that the high clocks had to be planned from the very inception, knowing that the 14nm LPP process delivered an increase in fmax.
Watching him post now is just constant damage control. Desperate to hide the fact that hes talking out of his ass most of the time.

>>56193103
Who are you talking about?

>>56193171
The Stilt, semi famous overclocker and forum tech "expert."

>>56193103

It is impressive how he has been so consistently wrong recently about basically everything he chimes in on you could almost use him to predict whats going to actually happen just by knowing its the opposite of what he says.

Can't knock some of his bios mods though.

eu is a commie organization. they sued Microsoft for making ie the default browser of the windows. "anti-competitive" my ass. more like "i am the market manager"

>>56193305
Intel also lost in US courts, anon. Not just EU courts.

>>56193314
This. Intel clearly broke all kinds of laws, these laws are for supporting a free market. What Intel tried to do is make it a monopoly, and stockholme syndrome tards somehow think Intel does good for them by selling them dual cores for 150 bucks (kek)

>>56193314
only eu fined them, they had settlements in the us courts. eu took their money like a mafia.

Hope AMD can make some money/marketshare this time around, Intel's getting squeezed with ARM designs as well.

File: FB_IMG_1471695389725.jpg (34KB, 640x491px) Image search: [Google]

34KB, 640x491px

>>56180905
Kidd is right. I work at AMD and we get fucked by Intel engineers ever day.
Good thing is my asshole starched a bit, so it doest hurt that much.

>>56191829

You weren't even floating in your daddies nut sack in the late 90's were you?

>>56192225
Asian market doesn't care for Apple anymore.

>>56193512

They will.

>>56191829
I had a cheap Sempron Laptop that ran like a dream. KYS

>>56193103

He bought in to the hype of himself. Now that his egos been fed, he needs to keep it fed or fall back to obscurity.

He posts on reddit as lolfail9001

>>56193512
ARM isn't doing anything to the HEDT or mainstream computer market, it isn't even remotely close to capturing server or HPC market share either.

AMD will grasp some of the enterprise market by offering competitive throughput at a lower price point. 32c server parts are nothing to scoff at. 128 threads in 2 sockets is ridiculous.

>>56193548
My first pc was P2, dipshit.
>>56193625
>sempron
>good
Kill yourself

>>56193637

>lolfail9001

There is an account on OCN with that name iirc - could it be The Stilts alt?

>>56190361

>intel does shady things so it's ok for AMD to do the same

nice try shillbot

File: 21.08.2016 - 05.40.55.jpg (3KB, 101x64px)

3KB, 101x64px

I hope they can pull off a product that really offers a good alternative to Intel right now. But after the whole RX480 lyingclimax I'm not that positive about it.

Let's just hope they pull the Intel prices back down.

>>56192885
>They chose 3ghz so that no one could guess what their final frequencies will be.

I seriously doubt that. AMD is grasping for any advantage they can get, if a Zen 8 core CPU could clock to the same level a BDW-EP 8 core CPU could then they would show it and be marketing it as such. The fact that they aren't implies that Zen is having trouble clocking as high as Intel's chips, which is likely considering the weak process and AMD's history (they could only exceed >4ghz with bulldozer, phenom 1 and 2 before that struggled to >3.5ghz).

I hope for the best with zen, but if I'm disappointed I'm going to get a 6700T and an asus H170I-PLUS D3.

File: logan mooooore.gif (698KB, 445x479px)

698KB, 445x479px

>>56184173
>captcha ENTER casa

>>56193786
Pretty much this. When intel revealed core architecture they went all out and amdfags called BL and didnt't believe it until chips were shipped. If AMD could puss 4GHz+ I doubt they won't falunt about it.

>>56193786
>this same tech illiterate retarded 3rd worlder is back

The purpose of the comparison made is to show performance per clock of the architecture.
Each chips has entirely different turbo behavior, they aren't going to be left stock. They aren't going to overclock them either. They were both placed at a nominal frequency. Thats all there is to it.
Your tired FUD is so pathetic I can't even find the words for it.

And yet again, implying there is any issue with 14nm LPP is beyond stupid.

File: gwxmYSF.png (32KB, 627x280px)

32KB, 627x280px

Intel already knows that Zen will be a failure. That's why they're bringing back the overclock-crippling FIVR.

>>56193854
Cannon Lake is low power mobile only.
2018 desktop platform is Coffee Lake which ships with 6 cores.

Yeah, intel totally isn't worried about Summit Ridge under cutting their i7E line at all. They're closing the gap themselves for no reason.

>>56193852
>And yet again, implying there is any issue with 14nm LPP is beyond stupid.

Every part that's been fabricated on the 14LPP process so far has been a dud. It's pretty clear that Samsung went for better density as any cost and has made an inferior product as a result.

>>56193876
>Cannon Lake is low power mobile only.
>2018 desktop platform is Coffee Lake which ships with 6 cores.

that is only according to some bogus leaked slides, there isn't any indication that this will be reality.

>>56180905
My dad works at Nintendo btw

>>56193852
>And yet again, implying there is any issue with 14nm LPP is beyond stupid.

AMD is already abandoning 14LPP for their newer console parts in favor of 16FF. I'm sure if they weren't bound by the WSA they wouldn't be using the anchor that is 14LPP for poolaris and zen.

>>56180905
My son works at the NSA, they design their own chips, latest version is Snowdenleak 7GHz 20-core

File: 28-14.jpg (87KB, 1064x714px) Image search: [Google]

87KB, 1064x714px

>>56193894
Its clear you're too much of a tech illiterate subhuman to even understand whats being discussed.

The 14nm LPP process provides an increase in frequency over 28nm HPP at lower power. This is exactly what is shown in Ellesmere. The average RX 480 clocks higher than any prior GCN based card on average, and it reaches these clocks with less voltage than prior cards need to hit even 1000mhz. The relative performance of the GPU as a whole is a product of its design, not of the process its fabbed on. If it had 64 ROPs then it would out perform the R9 390X at lower clocks and power consumption. The die layout was designed to be cheap and yield high.

Carrizo APUs are fabbed on the exact same 28nm HPP process, and hit 3.5ghz with a 3.8ghz turbo.
There is literally nothing but tech illiterate retard FUD hinting at a clocking problem anywhere.

>b-b-b-but these OEMs talking about Coffee Lake are all bullshit!
Hilarious how you pick and choose what forum rumors to believe.

>>56193966
>The 14nm LPP process provides an increase in frequency over 28nm HPP at lower power.

Only according to your marketing material. We saw from poolaris that 14LPP only brought max frequency up by around 200mhz, from the previous max OC of 1150mhz on GCN 1.0/1.1/1.2 to 1350mhz on GCN 1.4.

>Carrizo APUs are fabbed on the exact same 28nm HPP process, and hit 3.5ghz with a 3.8ghz turbo.

With a CPU architecture that was designed to hit >5ghz with slow caches and other design features that accommodate higher clocks. Zen does away with this and we will likely see a return to Phenom II-era clocks where the chips topped out at 4ghz.

>>b-b-b-but these OEMs talking about Coffee Lake are all bullshit!

[citation needed]

The only news about coffee lake was a hilariously low quality slide, anyone could fake something like that.

>>56194015
Clock speed uplift is relative, you subhuman retard.
GCN has an 8 stage pipeline, 200mhz is a 20% uplift over 1000mhz. That is a huge increase in fmax from process alone.
There is no such thing as GCN 1.X either, though you just outed yourself as the one and only tech illiterate retard who shows up in these threads to shitpost and spread FUD.

>>56194051
>Clock speed uplift is relative, you subhuman retard.
>GCN has an 8 stage pipeline, 200mhz is a 20% uplift over 1000mhz. That is a huge increase in fmax from process alone.

you are extremely delusional if you think a CPU design will see anywhere near the same uplift a GPU design gets from a node shrink

>There is no such thing as GCN 1.X either, though you just outed yourself as the one and only tech illiterate retard who shows up in these threads to shitpost and spread FUD.

>subscribe to my marketing terminology and blind religious belief in AMD to do the impossible or else you're retarded

>>56194081
Yep, you're that exact same NEET retard.

File: It's a conspiracy!.jpg (58KB, 591x443px) Image search: [Google]

58KB, 591x443px

>>56194105
>Yep, you're that exact same NEET retard.

>anyone who debunks my marketing material is the same person out to get me

It's all a conspiracy against you and AMD.

>>56194130
>shitposting NEET thinks he doesn't stick out like the sore thumb the way he makes the exact same posts almost word for word when he shows up in every single thread

You probably have diagnosed autism too. Living off welfare, NEET?

>>56180726
>No performance figures
>No benchmarks
>No engineering samples
>Not a real review
>NO proof of any kind

Yep. Sure is the rx 480 all over again.

When will you chucklefucks learn?

>inb4 it releases and is wholly abysmal and AMD claims they meant it for a budget market all along.

>>56194156
You're probably the same pajeet eating amd shit.

>>56194187
I'm whiter than you, subhuman autistic NEET.

>>56189512
The Zen chip an engineering sample so it probably isn't as refined and isn't able to click as high as the final chip will be able to.

File: perfwatt_1920_1080.png (27KB, 500x610px) Image search: [Google]

27KB, 500x610px

>>56193966
>there's nothing wrong with 14LPP

you have to be legitimately retarded or the world's most delusional AMD fanboy in existence to believe this. It's plain as day that GCN needs a major overhaul instead of constant rehashed and that AMD needs to get out of the wafer supply agreement if they want to compete in the future.

>>56194249
If amd changes gpu arch GCN fags are left with no drivers. ;^) Just ask the pre GCN card owners. ;^)

>>56182890
they should change that number nobody is going to buy a card called 490 retard

File: Polaris-10-die.jpg (885KB, 2048x1152px)

885KB, 2048x1152px

>>56194249
>I have literally no idea what I'm talking about: The Post

Ellesmere/Polaris 10 is a cheap, simple, 32 ROP design clocked high enough to be somewhat competitive with larger 64ROP designs. Pixel throughput is the only metric the GPU lacks in, otherwise its 36CU would be performing on par with the 44CU in the R9 390X. As amply shown in the chart already posted the die can hit 1000mhz at 0.85v~, a tremendous reduction in drive voltage for equal clocks to the rest of the line, power per transistor is reduced by a huge amount.
Its paired with 8GB of GDDR5 that draws 40w by itself, of the 150w power target the ASIC only takes 110w.

You're utterly clueless. Demonstrably so. You don't understand anything about process metrics, stop pretending like you're smart enough to be relevant in a conversation.

>>56194409
>As amply shown in the chart already posted the die can hit 1000mhz at 0.85v~, a tremendous reduction in drive voltage for equal clocks to the rest of the line, power per transistor is reduced by a huge amount.

Doesn't mean anything if you end up having to clock higher to compensate for the bad process.

>Its paired with 8GB of GDDR5 that draws 40w by itself

that's AMD's fault for not making proper power states for their memory, NVIDIA implemented frequency&voltage scaling for gddr5 so there's no excuse for AMD not to do the same.

>of the 150w power target the ASIC only takes 110w.

which is at reference clocks, where the 480 becomes so slow that even a 970 will outperform it. a 1350mhz 480 will draw closer to 250w.

File: 1471026902650.jpg (48KB, 627x626px) Image search: [Google]

48KB, 627x626px

>>56180726

>>56194330
*name

>>56194460
Stop trying, subhuman NEET shitposter.

Polaris 10 having 32 ROPs has literally nothing to do with the process its fabbed on, it is a design choice to keep the die layout simples. ROPs are bundled to back end of the CU.
You don't even understand what you're shitposting about.

>>56194559

Nobody is talking about ROPs here. We're talking about the shortcomings of 14LPP. AMD designing bad products (and rehashing the same design since GCN 1.0) has nothing to do with the fact that Samsung and GloFo have delivered bad process tech.

>>56194559
take your meds

I want AMD to be strong, but this currently just looks like hype building. AMD wants people to be hyped, and the "tech journalists" just want page views. NMS recently proved that clickbait and baseless hype is alive and well in journalism today, and people fall for it so these websites are encouraged to do it even more.

Come back with actual benchmarks, then we can talk.

File: download (2).jpg (4KB, 200x144px) Image search: [Google]

4KB, 200x144px

>>56194578
Anon pls.

>>56194578
>even more shitposting

Polaris 10 targeted clocks north of 1000mhz because it has 32 ROPs.
Per clock it has half of the pixel throughput of the Hawaii die which is a 64 ROP design.
This was a design choice.
AMD designed Polaris 10 as a cheap and simple ASIC which would yield high.
The higher clocks increase pixel throughput.

Not a single thing about the RX 480 indicates a problem with the 14nm LPP process.
Your shitposts aren't even coherent. Just rambling from a tech illiterate 3rd worlder.

>>56188839
>BRZ
GT86 reporting in

>>56194615
>Polaris 10 targeted clocks north of 1000mhz because it has 32 ROPs.

they wouldn't have chosen a process that tops out at 1000mhz efficiency-wise if that was the case

>Per clock it has half of the pixel throughput of the Hawaii die which is a 64 ROP design.

so AMD rehashed/rebranded their design? wonderful

>AMD designed Polaris 10 as a cheap and simple ASIC which would yield high.

p10 wouldn't be 232mm2 if that was the case. polaris 11 is what is designed for high yields as it's going to be in AMD's APU and mobile offerings, which is the segment they're currently experiencing the most growth in (as nvidia is abandoning it).

>Not a single thing about the RX 480 indicates a problem with the 14nm LPP process.

the 480 is a disaster that can't even beat Maxwell despite the new process. I seriously doubt AMD, even with their incompetent engineering team, would take such a large step backwards without the process being a part of the problem.

>>56194800

>the 480 is a disaster that can't even beat Maxwell despite the new process

Uising that logic one could argue maxwell is a disaster itself as baby maxwell still struggles to outperform hawaii.

>>56184069
Use 4chanX and image hover, dumbo

>>56194874
>Uising that logic one could argue maxwell is a disaster itself as baby maxwell still struggles to outperform hawaii.

what? hawaii has close to 1/2 the perf/watt of maxwell. it was far from competitive and was only a good buy if you could find a 290 for $200 or less.

>>56194891

> hawaii has close to 1/2 the perf/watt of maxwell

I made no mention of perf/watt - merely that the chips released are competing against a design at least a year older for raw performance.

>>56194800
>this tech illiterate retard is still trying

Peak efficiency in ops/watt comes down to the architecture.
The process and arch show a perfectly flat clock/voltage curve up to about 950mhz because of how FinFETs operate. Once the transistor receives enough voltage to operate it has a huge clock window, they don't perform like a planar transistor.
The uplift you start seeing on that curve does not mean thats where perf/watt ends.

The die is not designed to be as small as possible, the exact opposite is the case. Comparatively the Polaris ASICs are the least dense GCN designs ever. This was done to lower the defect rate, and increase net yield over all. Density was balanced vs die area. Lower die area means more candidates per wafer, but it does not guarantee more viable candidates.

Yet again, not a single thing about the RX 480 indicates a problem with the 14nm LPP process.
Your fanboy shitposting amounts to nothing. I deal in the realm of cold hard facts, autist. You should try it sometime. Then again if you were actually capable of rational thought you wouldn't be a worthless NEET drain on society.

>>56194919

and they require 2-3x the power draw to manage that, which also means that hawaii isn't competitive.

>The die is not designed to be as small as possible, the exact opposite is the case. Comparatively the Polaris ASICs are the least dense GCN designs ever.

then AMD shouldn't have done that because 14LPP has terrible area scaling. they probably would've been better off by fusing off more than just CUs like they normally do.

>>56193691

They're both Russian, both talk about f-state power curves. No idea though.

>>56195009
14nm LPP has ridiculously high area scaling, retard.
Its up to 15% denser than 16nm FinFET.
It shows an over 70% advantage vs 28nm HPP. If Hawaii were directly ported to 14nm with max density it would be a 130mm2~ die.

Stop talking out of your ass about things you lack the IQ to understand.

>>56195084

density doesn't mean anything if the process has terrible yields that destroy any design larger than ~100mm2. same issue intel is facing, forced to design smaller and smaller dies as they design new processes.

>>56195009
Just wait till Nvidia finally is forced to put back all the hardware they ripped out of their designs and their chips will for sure be using a lot more than the AMD equivalents

>>56195151

Fermi is my greatest ally.

>>56181641
THANK YOU, BASED JIM KELLER!

>>56194330
Yeah no one will buy a 1070 either

Or they might just call it Vega/Fury