HAHAHAHAHAHA OH WOW Vega might still be questionable but Intel

>>58348218
Also F3 stepping, apparently there's a F4 stepping with a 4.0 boost making the rounds.

This might be the 'black edition' Zens

S-stop..

>>58348218
My goddamn core i5 has 3.8ghz base clock and boosts up to 4, who the fuck cares? More cores won't do shit for AMD.

>>58348413
This is an 8 core cpu retard, comparable i7 6900k has 3.2ghz base and boosts to 3.7ghz

>>58348455
The 8350 is 8 core, and barely compares to an i5 6600k.

>>58348634
The cores function differently. 8350 cores aren't the same as i5 cores. Zen core architecture is more closely related to Intels now AFAIK.

>>58348754
>uploading the thumbnail

File: index.png (18KB, 130x97px) Image search: [Google]

18KB, 130x97px

>>58348787

>>58348787
ayyyymdfags can't afford proper internet

File: am4_confirmed_from_cryorig.png (44KB, 2010x641px) Image search: [Google]

44KB, 2010x641px

>>58348218
Also guys, FYI I got this conversation back from Cryorig last night...

Will there be a quad/hex core zen? I still use some programs that only support a couple cores and care more about core performance.

>>58348455
>Cores won't do shit for AMD

>>58348947
Why are you even here?

i need to see how it overclocks first on a good air cooler

>>58348947
Are you living under the rock?

>>58348218
Please be $500
Please be $500
Please be $500

>>58348634
did you miss the entire million and one conversations about IPC between generations and just go full pentium 4 with your response?

>>58350546
8c16t will be at least $500.
The fangirls thinking it will be ~250 are beyond delusional.

>>58348837
R1 Ultimate on AM4? perfect

File: 1467561083674.jpg (29KB, 320x283px) Image search: [Google]

29KB, 320x283px

>4Ghz boost
>95watt
if this overclocks well...

Zen should have open boost clock as well, so if the thermals allows it the CPU will increase the boost clock automatically.

File: 1483675225835.png (6KB, 65x49px) Image search: [Google]

6KB, 65x49px

>>58348811

>>58348947
Quad core Zen is a given, a few months after 8 core Zen.
We still don't know about 2 or 6 core Zen because of the Core Complex thing which always includes 4 cores.
Maybe they have native dual core SKUs ready, or alternatively will cut down defective 4 or 8 cores.

>>58348218
Does this really matter?

If youre gonna spend hundreds of £/$ on a CPU then chances are you'd overclock it past 4GHz anyway.

>>58348218
So if the Zen at 3.4ghz they benchmarked earlier showed them to be about even with Intels 6900K that was running at stock boosts then at 3.9 Zen would be ~15% better or in other words seven Intel generations ahead. RIP

File: 1483674966232.jpg (170KB, 857x595px) Image search: [Google]

170KB, 857x595px

>>58348811

>>58353095
~15% better at single core and at multicore, this will double/triple to maybe 30-40%

>>58353095
Depends on the workload. AMD showed only Blender. That French magazine with the 3.15/3.4 ES chose a wider range of benchmarks and the 6900k was about 10% faster in gayming and 14% faster in rendering/encoding.
Zen at 3.6/4 could be slightly ahead of 6900k, but not much.

It's all guesswork and within margin of error anyway.

There's no way GloFo would be able to make a 4.0GHz chip, simply impossible.

>>58353196
Even if Zen is just barely equal to the 6900K, that's better than AMD has managed since Phenom

>more cores

I just want IPC, more cores is useless!

So what Zen CPU is going to be ~$300?

Is it this 8 core i've been hearing about or is that $500+ bullshit no one will buy?

>>58355056
>true 8 core
>$300
You must be retarded or 12yo. FX8xxx are not true 8 cores.

>>58348218
hahaha oh wow
16 amd threads = intel's 2 physical cores with HT

Eat shit ayymd cancer. Stop posting on my board

>>58348218
Single thread performance, how good it OCs and how much it costs. Then we are talking. I'll be fine with even two cores, provided they are good and cheap.

>>58352693
>>58353150
>>58348811

i hate you all so much

>>58355087
I never said they were?

I said ZEN not FX8xxx, I don't give a fuck about a 4+ year old CPU.

I want to know how much the Zen 8 core mentioned in the FUCKING op will cost.


Retard

>>58355121
woah..that edge

>>58352724
There will be 6 core SKUs, quad cores are actually unlikely at this time.
To make a 6 core chip they disable one core per CCX, to make them symmetrical. Each CCX loses one core and one L3 block.

>>58355136
whoa, that retardation

No one even mentioned the FX8350 or any other FX series CPU, so why the fuck would you even mention it?

>>58355154
FX costed $300 because they weren't true 8 cores. You can't expext zen 8 core to cost the same.

>>58355176
So you're saying it costs as much as a 6900 and performs the same or worse?

Wew lad, AMD is retarded again.

I bought a 6 core haswell-e over a year ago for $300. Sounds like Zen will be similar per clock performance as haswell-e, but will cost basically the same as I paid ~14+ months ago?

>>58348218
>...Intel is confirmed slaughtered

Errr, no prices are released, senpai.

It's nice they finally made a fast CPU* but we'll have to see if they can compare to Intel price wise.

* if they didn't cheat the demos.

File: bd module.jpg (60KB, 704x412px) Image search: [Google]

60KB, 704x412px

>>58355087
Yes, yes they are.
Two integer cores per module, with a shared FPU.
A shitposting autist here named Tony Dickey tried to sue AMD over this, and the case was thrown out.
A US court decided that a module is in fact two cores.
Deal with it.

>>58355199
US courts decide a lot of things that are objectively incorrect

>>58355301
Except they are objectively correct here.
There is no standardized definition of what constitutes a "real" core anyway, kid.
The module has two integer cores. Deal with it.

>ayymad overclocks to 3.9 ghz
>intel does 4.5 stock

HAHAHAHAHAHAHAHA

>>58355301
>no matter how high the authority I will dismiss it in favor of my own shitty opinions
you weren't called stupid enough as a kid

File: 1458923695936.png (163KB, 380x403px) Image search: [Google]

163KB, 380x403px

>>58355327

4.5 stock a 8/16 core, you dumb fuck?

Go back to >>>/v/, manchild.

>>58348218

>Windows 10

Found the problem.

File: IMG_1993.png (100KB, 840x630px) Image search: [Google]

100KB, 840x630px

>>58355356
I PUSH MY FINGERS INTO MY EYES

>>58355382
thats a 4 core 8 thread CPU at 4.8GHz

do that with a 6900k and report back ;)

>>58355398
Moving the goalpost, SHILL!

>According to the report from TechReport who got to thoroughly test the Vega sample with 8 GB VRAM in DOOM, this specific sample is somewhere in between the GTX 1070 and GTX 1080 levels of performance. When set to Nightmare settings, the game reported max frame times of 24.8ms. We also recorded some gameplay of DOOM running on the Vega sample while we were at AMD’s CES event.

>not even released
>already lost to 1 year old Nvidia cards

HAHAHAHAHAHAHAHAHAHA

Kek even a pentium 4 can go faster than Zen

Zen confirmed wreked

>>58355403
How is that moving the goal posts, also i'm not that other poster.


He said 8 core 16 thread, you posted a 4 core 8 thread, you're the only one being retarded here buddy.


Also, posted from my 6 core haswell-e Fuck yourself.

Ayymd BTFO again. What a surprise

>all these Intelfags wishing doom on AMD
It's like you enjoy being raped by a monopoly

>>58355435
Just accept your defeat with your Pajeet Rayjeet CPOOs

You can smell the fear here :^)

>>58352275
I'm thinking they'll have the lower end 8c/16t Ryzen at ~350 and then just go up from there. If you go based on their k10 chips where they were 1/3 the price for the same performance that is. But even at 500$ they'd be 1/2 the price and leave plenty of room to undercut Intel if they try to compete on price point. Hopefully they don't go beyond that, though by every right they probably could.

>>58352433
The French magazine writer left a binary string easter egg in his article that translated to zen=5g@air. He was later asked about it and confirmed he overclocked his engineering sample to 5ghz on a single core using an air cooler. The real question is how well / how far their dynamic boost will go. That msi platinum board has an 8+4pin cpu power connector which would suggest it responds well to loads of free power.

>>58353196
He was running 3.15ghz on multi threaded applications, and 3.5 on single threaded, 3.3 on 4 cores. It had a ~10% clock deficit.

>>58355435
We don't want AMD to suck, they just do. No amount of my hoping AMD does better is going to change the fact they suck tits.

if they release a good CPU, fucking great for them I hope it sells well. But I'm not going to shill garbage for them because they're poor and can't compete with intel.

>>58355469
>zen=5g@air.
As I recall he also disabled 7 cores.

>>58355435
We're wishing doom on dumb zenposters, not zen

>>58355479
Celerons hit 5GHz on air

>>58355435
>arguing apples to apples

They're both NDA locked down, untrustworthy x86.

>>58355482
Most dumb zenposters are just people messing about.

By the way, intel is finished, hold onto your jew hats because you're gonna have to sell them when you're BANKRUPT
Nazen master race

File: ASRock X370 Taichi.png (1MB, 800x942px) Image search: [Google]

1MB, 800x942px

>>58355469
There are some Gigabyte and ASRock AM4 boards coming out that are absolutely fucking gorgeous in power delivery. A couple MSI boards are pretty nice too.

>>58355479
Yes, it was single core. He stated that the board's VRM was too lacking. It was a bare validation board.

>>58355516
>blaming the electricity for poor OC performance
It's like audiophile power lines

>>58355560
This is getting silly now.

>>58355560
It's like telling people they shouldn't cheap out on their PSU

>>58355403
holy shit you are retarded

File: GA-AX370-Gaming5.png (688KB, 761x916px) Image search: [Google]

688KB, 761x916px

>>58355560
Yes, anon. Thats how overclocking works.
If you have only a few VRM phases, if they have bare mosfets, they will overheat and throttle.

Thats why high end boards have many phases with heat sinks on the mosfets.

>>58355575
Stop it.
Nobody takes your shilling seriously

>>58355575
Blaming the motherboard for being unable to get a stable overclock on 1/8 of the CPU is kind pushing it though

>>58348734

Ow, sure mister know it all. Can you take another peek into your crystal ball and tell us more about a CPU that DOESN'T EXIST YET.

>>58355479
they also claimed that the motherboard's VRM was the limiting factor

>>58355495
piledriver hits 5ghz on air and it's terrible

zen's clocking capabilities are only relevant here because we know it likely has broadwell-e tier IPC

>>58355479
>He was later asked about it and confirmed he overclocked his engineering sample to 5ghz on a single core using an air cooler.

>on a single core.

Reading is hard.

>>58355516
Yeah, I like how manufacturers are taking advantage of the largely bare PCBs to do custom prints and what not. Though a lot of them look kind of fucking stupid.

Didn't catch the part about the test boards lackluster vrm. Good to know. As it stands I'm sitting on a 4.3ghz e5-1650 v2, so unless it's just an absolute ass blasting I probably won't upgrade until zen+.

File: 1289330658631.jpg (32KB, 400x400px) Image search: [Google]

32KB, 400x400px

I love how certain camps aren't even trying to discuss anything now and just shitpost to bury other posts.

Delicious, it's almost like someone's bleeding from their ass all the way to their hospital room

>>58355588
No its not.
The board they had was a validation board. Literally the bare minimum of what is needed to check and see if everything boots.
They did overclock all 8 cores, and managed to hit 4ghz with their set up.

>>58355603
>because we know it likely has broadwell-e tier IPC

NO IT FUCKING DOESN'T AND YOU KNOW IT

>>58355516
>>58355575
>shitty PCIe placement on both, can't get 2 slots of room between 2 graphics cards
I don't understand why so few manufacturers manage to get this right.

>>58355603

>broadwell-e tier IPC


Ow man, my sides.

Uhhh UUUU how can AMD do this to us, It's not fair to us who bought a 6900k

>>58355618
>>58355636
>in spite of the overwhelming evidence I choose instead to believe in my own comforting bullshit
the intelfag manifesto

>>58355608
i don't like it. it's very repetitive, boring and unfunny.

>>58355635
But anon, the first PCIe port has a 3 wide gap with the second.

>>58355604
I'm not crazy about all the graphics they're throwing on there, but its a rare sight to see an AMD socket board with beefy VRM and good features.
It is super apparent that NB+SB on die really does open up a lot of space on the board though. If you take away the M.2 slots, and other add in features then a lot of the board would be bare. I want to see the X300 mITX boards they'll come up with.

So far they've only really detailed the entry level APU boards, and these high end X370 boards.

>>58355636
CanardPC explicitly stated Zen has near Broadwell IPC.

>>58355660
>overwhelming evidence
AMD keynotes hardly consist of 'overwhelming evidence'.
I'll believe it when it's out and reviewed by third parties.
Until then it'll always be a 'maybe' no matter what the claim.

File: ZenTests.jpg (212KB, 708x733px) Image search: [Google]

212KB, 708x733px

>>58355636
Yeah, sorry, he meant slightly higher than BDW-E.

It's a good thing Zen can't go higher than 3.1 or Intelfags might have to kill themselves!

>>58355650
correct. such things should rightfully cost over a thousand dollars because they are difficult to make and very good products. AMD is devaluing the hard work of Intel's hard working, highly educated engineers with its immoral product of sorcery.

Jim Keller should be jailed for his crimes against capitalism.

>>58355666

in between the GPUs he means, idiot.

>>58355665
But isn't watching these underage children run around replying to every positive post with pure, undiluted shitposting amusing in its own way? Imagine a time when these will be replaced by trained apes or just by if then if else scripts?

>>58355696
That made no sense so I assumed otherwise.

>>58355676
but shill-kun, it has already been reviewed by some french faggots

>>58355635
One gap of room between cards is enough

>>58355705
They didn't test Cinnebench of SuperPi so it's not relevant.

>>58355708
but muh airflow

File: MSI Z270 Gaming 5 review.jpg (71KB, 590x332px) Image search: [Google]

71KB, 590x332px

>>58355575
>>58355516
Look at how shit and barren those motherboards look like in comparison to Z270 boards, especially ones that will sell for a similar price.

Not even the motherboard manufacturers expect Ryzen to sell at all, so they've intentionally held back their best designs and features for superior Intel platforms instead.
A
M
D O A

>>58355666
And how many free slots will you get between the actual cards if you install standard models with 2-slot coolers?

>>58355708
If you want higher temps and louder fans on the top card, sure. But that's stupid when it can be avoided by something as basic as intelligent PCIe slot placement. No gaps are "enough" too, the cards will work without bursting into flame, that's good enough right?

>>58355705
Yes, it counts as evidence but it's preliminary evidence.
You now have reason to say it'll 'probably' be whatever you claimed (I cannot remember what claim we're arguing about) but not certainty. The consumer release is not what they are displaying, and for all you know the consumer release might be better than the engineering samples, so it's best to not commit such evidence to 'hard truth' just yet.

why the fuck are AMD not releasing clocks and pricing publicly already

also this rumor about no four/six core better not be true. i dont care how cheap the eight core is, not having a cheaper four core solution ready to go at launch is bonkers

>>58355676
>t-they must have faked it somehow... how can it be
look buddy, I said "likely". it's also possible those few benchmarks we saw were outliers, but you can't exactly call handbrake and blender "synthetic".

in light of the other leaks we've seen, nothing suggests that IPC isn't either on par or very close. right now the only thing that can be reasonable doubted is whether or not the 4c/8t zen chips will be able to clock high enough to match the i7s, and if it makes you feel better, they probably won't.

>>58355704

Well, I guess that makes sense because that's the space relevant for cooling.

File: 1421541090157.jpg (46KB, 512x699px) Image search: [Google]

46KB, 512x699px

>>58355715
>tests that don't conform to my arbitrary conditions aren't relevant

B U T H U R T
U
T
H
U
R
T

>>58355760
That's no rumor, it's fact. AMD is launching high-end first.

>>58355715
and when somebody does and you don't like the results what else will you pull out of your ass?

File: GDyOS[1].png (224KB, 778x960px) Image search: [Google]

224KB, 778x960px

This is still relevant today

>>58355741
Many blower card designs have vents in the PCB for card-against-card placement

>>58355780
source?

>>58355760
They officially stated that at launch chips would have base clocks speeds of 3.4ghz or higher. So far they're considerably exceeding that on the high end. They're probably not making any definitive statements just yet because they keep hitting higher clocks in these recent steppings.
Price is very likely in the range already leaked. $500 absolute top end SKU. A lesser 8 core chip at $350. Lower core counts below that.

There will be 6 core SKUs. They disable one core and L3 block per CCX.
Not sure about how they'll handle quad cores, CanardPC said on twitter that they'll be coming later.

Either way the 6 core chips should be very affordable.

>>58355760
>why the fuck are AMD not releasing clocks and pricing publicly already
releasing unfinished hurts them more than releasing early.

>>58355795
>it will increase performance 1%, thus beating AMD by -85% and increasing our sales to holy shit intel went broke ive lost my job

ftfy

>>58355795
I'm ready for the updated version
>it will increase performance by 1%
>this beating AMD
Just how badly AMD gets beaten is the only question left

Haven't bought AMD in 15 years but I want to see them succeed again. Hope Zen is gud

File: dan_dobberpuhl_jim_keller-380x285.jpg (30KB, 380x285px) Image search: [Google]

30KB, 380x285px

Did you doubt the man that was involved in making histories coolest CPUs like K7 and K8, Dec Alpha 21164 and 21264, A5/A4? Involved in making Hypertransport and x86-64? This guy won't stain his name by making something shitty.

>>58355795
Intel HQ
>What do we do after Skylake?
>LEL INCREASE CLOCKS 200mhz AND PRETEND ITS A BRAND NEW CHIP INSTEAD OF A MINOR UPGRAD LIKE DEVILS CANYON
>but reviewers won't like that
>I'M BRIAN KRZANICH AND I TELL PEOPLE WHAT TO LIKE
>b-but sir what do we do for the desktop market after Kaby Lake
>ADD MORE CORES TO THE EXACT SAME CHIP AND SELL IT AGAIN


Intel has a 10 core HEDT chip.
Intel has a 6 core mainstream chip coming.
AMD has an 8 core HEDT chip.
Intel is literally the moar coars company. AMD is increasing IPC, Zen+ will bring another big IPC uplift as well, and on a brand new 7nm SOI based FinFET process while intel is still puttering on bulk silicon.

>>58355796
Many? Vents in the PCB? Like what, the fucking 9800GX2?

>>58355798
http://arstechnica.com/gadgets/2016/12/amd-zen-performance-details-release-date/
It's common knowledge basically, you can find many more articles with the same information.

>>58355883
prepare for a year+ of AMD shills pretending they're just as good as intel until intel releases a new core architecture in 2018 to BTFO them again.

>>58355883
Intel has a 72 core chip, you can't blame them for enveloping the entire CPU market in its entirety, they probably got bored of blazing fast couple-core chips and needed something to sell while they make non-silicon CPUs happen
By that time AMD should be fully caught up

File: Intel-Coffee-Lake-Roadmap.jpg (338KB, 1489x597px) Image search: [Google]

338KB, 1489x597px

>>58355904
>intel releases a new core architecture

Wow, a 6 core Kaby Lake? Amazing.
Bet it won't burst into flames either.

any news on the zen apu?
I heard it will be like a mid end gpu

>>58355904
JUST
WAIT
(™)

>>58355904
Hahaha. Nope.

2018 desktop is Coffee Lake. Same IPC as Skylake/Kaby Lake but with 6 cores.
2019 is Cannonlake for desktop. No major IPC uplift, just lower power.
They already have 3 new generations of arch laid out, and two of them are just 10nm refreshes.
Intel has nothing new coming out until at least 2021.

Intel will be on 10nm bulk for mainstream desktop and HEDT into the 2020s.
AMD will be on 7nm SOI FinFET in 2019.

Sad.

>>58355835
I don't doubt the man. I just think the science is too advanced for any individual to pull a K8 tier victory out of their ass again.

>>58355883
>AMD has an 8 core HEDT chip.
Zen is supposed to scale up to 32c/64t server stuff. They will put out 10+ core HEDT chips if there's demand for it.

"moar coars" was always a shitty /v/ meme anyway.

>>58355883
But that's what they said from the start. They even specifically talked about breaking their product release cycle to make a refresh of skylake, going away from dieshrink-archchange and moving towards dieshrink-archchange-optimization. They said it was skylake from the start
Cannonlake should be something different at least

>>58355952
>I don't doubt the man. I just think the science is too advanced for any individual to pull a K8 tier victory out of their ass again.
There won't be another K8 but he can get close, and so far Zen is getting close if there's retail SKUs with 4.0 boost.

File: image.jpg (42KB, 365x500px) Image search: [Google]

42KB, 365x500px

>>58355944
Did it look like Kanye was gonna win the election when you were in 2019?

>>58355924
>Intel has a 72 core chip
And it has Atom-tier cores and can't even do hardware virtualization, which is absolutely necessary in enterprise.
I wonder who would buy that shit.

>>58355835
You mean the same man who jumped ship in 2015 once he saw how badly Zen was turning out to be?

File: 1481935316973.jpg (65KB, 520x442px) Image search: [Google]

65KB, 520x442px

>>58355944
>AMD will be on 7nm SOI FinFET in 2019
>AMD shills actually believe this

>>58355959
>>58355959
>Cannonlake should be something different at least

In power usage, yes. But not for performance, look at the core sizes between Broadwell and Skylake, Skylake has a wider core with much more resources on all fronts and it's at best 5% faster than a Broadwell core, there's a clear showing that these cores can't be fed properly and Intel is planning on a new x86 architecture by 2020/21.

>>58355928
those will be out much later this year but even the older gen bristol ridge APUs look promising with the speed boost from the DDR4
https://www.youtube.com/watch?v=lGuFVc8AFXk

>>58355993
>anon doesn't know GloFo is skipping 10nm altogether

File: image.jpg (79KB, 705x497px) Image search: [Google]

79KB, 705x497px

>>58355998
Dis Kanye West win the presidential election in 2020?
tell me now, great wizard

File: 1419379215129.gif (249KB, 255x111px) Image search: [Google]

249KB, 255x111px

>>58355989
>implying
AMD only "rented" him to lay the base of the chip and then he went on to do something else.

But please, don't let the facts spoil your buthurt.

>>58355944
Intel has the means to make drastic changes to their long term strategy. If even AMD could survive on failed products for half a decade while they scrambled to make something good from scratch then you bet your ass Intel will.

>>58355952
>They will put out 10+ core HEDT chips if there's demand for it.
They won't, socket AM4 only supports two memory channels. There are no MCMs going to AM4.

>"moar coars" was always a shitty /v/ meme anyway.
No, its been around forever. Intel shills were spouting that on forums when AMD released their first dual core chips. They were doing it when AMD released their first quad cores. They did it when AMD released the 6 core Phenom IIs as well. It has been going on nonstop since the release of the Bulldozer arch.

Now intel is taking the crown, releasing chips with moar coars since they've stalled in increasing IPC. Now AMD is making huge strides in IPC.

>>58355959
>intel said they're now the moar coars meme company from the start
So good of you to admit it.

>>58355993
Cry more, anime poster.
They appended the WSA with GloFo to utilize it. GloFo will not offer a 10nm node, they will put IBM's 7nm SOI into production. Work on its ramp starts this year.
Poor little intel won't keep up.

>>58355998
Aren't these cores still loosely based on the Pentium Pro? AMD with Zen should have more breathing room since it is made from scratch.

>>58356021
>Intel has the means to make drastic changes to their long term strategy
But Intel's problems are short-term, and with its current joke of a CEO I can assure you such a massive comany isn't gonna change its roadmap overnight.

>>58348218

Oh boy, can't wait to buy one of these, what with all of the programs and games where 8 weak cores will outperform strong 4.

Just kidding, bought a cheap used 6700k after Ryzen was confirmed to be another MOAR COREZ ayymd placebo and kaby turned out to be nothing more than polished skykek.

File: image.png (437KB, 541x538px) Image search: [Google]

437KB, 541x538px

>>58356022
>skylake=more cores
>implying more cores is even bad with today's apis
I mean one of the biggest hypes of Kaby lake was the pentium...

>>58356012
they can name it whatever they want, their pitch measurements don't lie however.


Intel's 14nm is like 20% smaller than GloFo's 14nm.

File: 1476814520433.jpg (204KB, 922x882px) Image search: [Google]

204KB, 922x882px

>mfw ordered 7700k last night

>>58356053
Yeah, Intel's fabs are better, they'll need them when their arch is stagnating.
GloFo/Samsung/TSMC are moving faster than Intel, they have never been as close as they are now, remember that Intel used to enjoy a 2 gen process lead against all of them? Now it's barely half a gen.

At least it's not a situation where AMD was expected to compete with 28/32nm Planar vs a 2nd gen 14nm finfet.

File: 1417821812256.jpg (26KB, 233x188px) Image search: [Google]

26KB, 233x188px

>>58356045
>8 weak cores
keep telling yourself that

>>58356053
Intel's 14nm was soo shitty and buggy that its taken them years to iron out. They lost so much money that the have to keep using it through 2017 and into 2018 to recoup their losses. Kryzanich literally lied to investors about their foundry yields. They're beyond desperate.
10nm will be even worse.

Samsung's 14nm process was mass producing 80mm2 ARM SoCs immediately after it came online, and they were selling at a profit in a high volume low margin market.

It doesn't matter if intel produced a denser BEOL process, they objectively failed in the measures that matter for a foundry.

>>58355667
Yeah, the x300 is interesting. I like that they actually made a seperate X series chipset for mitx. Help keep cost down as there is no need for 32 lanes of PCIE. Though iirc that's the only real difference between the x370 and b350. So unless I'm missing something I don't see why they wouldn't just use the b350 chipset.

>>58355760
All information shows that 4c skus will be their new APUs. There will undoubtably be 6c skus, just a question of how soon they validate enough cut down 8c cpus to consider selling them.

>>58355835
Unfortunately all he really did was oversee the initial design. But I highly doubt he would put his name on something he didn't think he could make happen.

>>58355924
A functioning silicon-germanium alloy 7nm chip has already been manufactured, and is most likely what zen+ will be made with. Meanwhile, Intel is still fucking around trying to get 10nm on silicon.

>>58355993
https://www.extremetech.com/computing/235842-globalfoundries-announces-new-7nm-finfet-process-full-node-shrink

>>58356045
So you're suggesting intel's 8+ core skus have weak cores as well?

>>58356053
Doesn't change the size of the transistors themselves, just the spacing. By your numbers, The 7nm process would still be smaller than a 9nm intel finfet.

>>58356022
>They won't, socket AM4 only supports two memory channels. There are no MCMs going to AM4.
that's dismaying. I guess we'll have to get whatever platform Naples will run if we want a true X99 counterpart.

File: image.png (358KB, 552x543px) Image search: [Google]

358KB, 552x543px

>>58356082
>now it's barely half a gen
>it's already Kaby Lake vs Vishera
TIMES RUNNING OUT AMD, TICK TOCK TICK TOCK TOCK REBRAND TICK TOCK TOCK

>>58356076
is that a pelt or the cloud of billowing smoke from your 7700k?

it'll keep you warm at least

>>58356098
>Doesn't change the size of the transistors themselves, just the spacing. By your numbers, The 7nm process would still be smaller than a 9nm intel finfet.
you're assuming it continues the ~20% route, for all you know their 7nm is 40% larger than intel's.

You literally have no idea and are guessing.

>>58356098
What is the zen+ core count, clock speed, size and release date, foreseer of technology?

Honestly I'm more looking forward to Zen+, if.. If Zen is BDW-E IPC and Zen+ is expecting another 10%, then it should easily pass Kaby Lake's IPC.
And more importantly the clocks will be massively improved with a mature process.

>>58356133
By the time zen+ comes out Intel will be well into selling cannonlake, at least

>>58356045
>what with all of the programs and games where 8 weak cores will outperform strong 4.
Confirmed doesn't do any real computing

>>58356098
you seem knowledgeable enough. is there any credibility to the rumors that zen/polaris(etc) APUs will support shared HBM2 RAM?

>>58356144
Not according to their roadmap, ULV cannonlake is expected by the end of 2017 and desktop models by middle 2018 IF(and that's a big IF) yields aren't awful.

And cannonlake is a shrink of skylake, what exactly are you expecting from it? Intel doesn't do new arches + shrinks at the same time unless they want their yields to be 1.7

http://semiaccurate.com/2016/09/26/globalfoundries-7nm-process-isnt-even-close-name/

:^)

>>58356196
SOI FinFET vs electrostatically inferior bulk FinFET.
BEOL only influences area scaling.

Intel's foundry is now noncompetitive.

I wish mods banned every Intel/AMD thread right until Ryzen release. These threads are pure pointless shitposting.

>>58348218
>he takes a photo of his screen with a camera rather than taking a screenshot

typical windows user

>all this shit posting
>Product hasn't launched and there aren't any benchmarks to compare

>>58356229
That was took from the CES event, AMD won't let you take a screenshot and upload it to your phone.

>>58356223
we had a pretty good run actually, until you showed up

>>58356246
>Product hasn't launched
there's still a lot to talk about

Zen is doing some pretty brash shit and regardless of the outcome AMD's decision making process is deeply interesting

I hope you realize you're making the thread, board and website worse with your bitching

>>58356246
>there aren't any benchmarks
Is this how you convince yourself?

my next cpu and gpu will be AMD.

File: 1fxSexs.gif (694KB, 500x500px) Image search: [Google]

694KB, 500x500px

>>58356223

>>58356272
It's worth discussing and I'm curious to see how it pans out. Obviously, it performing well would be good for everyone as Intel and Nvidia would have to go back to more competitive pricing. It's just annoying how these threads are 90% endless shitposting. Like >>58356287

>>58356180
Cannonlake is a brand new arch you have no fucking idea what you're talking about stupid retarded AMD poorshill

>>58356196
>What GF is saying is that their 7nm process is a real, true, honest shrink but not from a real, honest, true 14nm process, it is from a real, honest, and true 20nm process. What do we know that as? 14nm.


Kek, GF 7nm is actually basically the same as intel 14nm

let me know when they show off doing ps2 emulation without overclocking then we can talk not that ill buy AMD again anyway

>>58356370
It's okay at least chips with GF7nm won't be (previous arch) x 1% :P

>>58356223
>I wish mods banned every Intel/AMD thread right until Ryzen release
They wont do that, but I had a janitor attempt to ban me earlier for posting a rice cooker thread.

Fucking jews keep 5+ of these identical Ryzen threads alive, but a single rice cooker thread triggers them.

/g/ is dead

>>58356399
>ignores the power advancements, more refined turbo states, and massive iGPU increases

Maybe, just MAYBE they weren't going for pure IPC gains.


People who don't understand technology should leave.

>>58356422
he just wants to live out his weird fantasy of a gaming PC without a discreet GPU

>>58356407
Like it was ever alive to begin with.

>>58356346
Sure. Skylake is "brand new arch" providing a whopping 2.7% IPC uplift over Broadwell.
I'm sure intel is really going to bring out the big guns again. Might even see a whole 2% uplift!

Intel's 10nm mobile Cannonlake launch is going to look exactly like their 14nm launch of Broadwell CoreM. Horrible yields, even on tiny parts. Literally dozens of SKUs because clock binning was so wild that they couldn't validate enough chips to hit the same base cocks. 5w TDP parts pulling 12w under load. Desktop Broadwell delayed. Lying to investors about foundry performance. One week later admitting theres an issue, but claiming you'll have it fixed in a quarter. Desktop Broadwell delayed again. Desktop Broadwell finally released, but only 2 SKUs, production immediately canceled, losing significant money on every chip. Skylake rushed to market, still mobile launch only at first. Took months to get desktop chips out. Skylake desktop chips impossible to find in stock for month after reviews released due to yields of high clocking chips still being abysmal.

Intel's 10nm parts will be a non stop laugh riot disaster. Kike Kryzanich is going to step down.

Don't even make me mention the billions of dollars intel lost on their ARM competitor Atom chips, their poor performing failed modems, and their piece of shit failed Sophia SoC.
Intel is literally falling apart internally and to top it all off they just fired 12,000 Americas to replaced with street shitters.

Intel legitimately is closer to falling apart and spinning off divisions than AMD.

>>58356454
That's the point though, a thread that is getting little traction will die on it's own in an hour at the most, but a janitor still felt it was his patrotic duty to the /g/ gods to report the thread to admins and try to get me banned. Yet the 5th or 6th Ryzen thread on this board gets a pass because we obviously need more of those.

Like I said, /g/ is dead.

>>58356422
Kaby Lake uses more or equal power to Skylake, speedshift is Wangblows only and is only useful on ULV/ULT/CoreMlol crap and the GPU is barely 12% faster.


It's a joke.
This looks more like a new stepping at best than something new.

>>58356396
Mom won't buy you another computer until you make at least a C in English anyway.

File: image.png (442KB, 500x712px) Image search: [Google]

442KB, 500x712px

>>58356463
Last year AMD turned a profit for the first time since 2009
How can Intel even compete?

>>58348218
AMD please protect my smile

>>58356486
B-but m-muh 4k Netflix, goyim.

>>58356519
What does AMD's situation have to do with Intel's internal issues? I thought you guys told me Intel is too big and strong to be affected by puny little AMD ;)

>>58356509
no im not spending enough for another motherboard and processor this soon after i have just done that because my fx 8320 was a massive and horrible disappointment ill stick with my 6700k for a few years at least and no it doesnt matter that im on an HDD

>>58356558
Yeah but then you came and told me differently I guess I lost

>>58356519
Nice deflection.
10nm Cannonlake will be a disaster. Intel already *knows* it will be a disaster.
Thats why their mainstream desktop platform for 2018 isn't Cannonlake, its 14nm Coffee Lake.
You won't see desktop Cannonlake until 2019, and thats if it isn't delayed numerous times.

14nm in 2018.
10nm in 2019.
10nm in 2020.

>>58356559
Nice blogpost.

File: ticktockticktock.png (394KB, 1011x292px) Image search: [Google]

394KB, 1011x292px

>>58356127
I know, just being a smartass. The pitch could be even more drastic than that. Could actually have less pitch as well seeing that they've changed materials.

>>58356132
Not a clue, anon. Just taking the info that's been out there and speculating. It is most likely that zen+ will include a die shrink. But not necessarily guaranteed.

http://www.digitaltrends.com/computing/amd-skipping-10nm-for-7nm/

>>58356168
I haven't seen anything on that, just people on here dreaming and speculating (myself being one of them). It's a possibility they might go balls out with something like 8-16GB shared HBM2 which would be amazing. I seen a gamersnexus video yesterday going over AMD's vega slides where they detailed a unified memory bus. I'll have to try and find more info on it though.

>>58356346
Skylake was the tock, anon, and intel moved to a tick-tick-tock developement model. Kaby makes one tick Canon/coffee makes two ticks. Canon and coffee lake are also die shrinks, which Intel has never done architecture updates on a node change in 10 years. What makes you think they would change this now, especially with their new dev model not lining up with it? Source me, please.

http://arstechnica.com/information-technology/2016/03/intel-retires-tick-tock-development-model-extending-the-life-of-each-process/

>>58355589
you need to chill out my guy

>>58356619
Kaby was supposed to be the "optimization", an excuse for rereleasing skylake, which would put cannonlake in line for the die shrink
But broadwell was also an unofficial tock, they've already been pulling this tick-tock-rerelease shit for a whole cycle at least

>>58353085
Yeah, the post OC stats will be the most telling

some AMD products barely OC at all

>>58356619
>Skylake was the tock

Correction, broadwell was the tock to be more accurate.

>>58356643
Yeah, the tock lands on their E platforms. Haswell tick, haswell-E tock, skylake tick, broadwell tock, kaby tick, (new model) Canon tick.

>>58356653
>some AMD products barely OC at all
Name me two architectures of CPUs that can't overclock at all.

Please don't confuse GPUs with CPUs, thanks.

>>58355102
We already know its single-threaded IPC is right up there with intel's current stuff, we also already know it overclocks well with all the headroom and features they're making for overclocking and the low power consumption of the part, naturally, helps in this area as well.

All we're lacking is REAL pricing information, but if the highest-clocked 8c/16t part is at or below 800 USD, we should be golden, because the 6900k is a good 300 USD above it.

...and let's not forget all the new features the CPU brings, as well.

>>58356688
He did say AMD products

File: 1474048854009.png (105KB, 255x345px) Image search: [Google]

105KB, 255x345px

CERTIFIED SHIT WRECKER

>>58356688
shieeet

File: certified shit wrecker.jpg (13KB, 381x286px) Image search: [Google]

13KB, 381x286px

>>58356715
Goddamn right.

>>58356146

If you do, that's cool and Ryzen will probably be great news for you, but the people who meme about Ryzen the most are usually just gaymen too wrapped up in synthetic test results to realize none of that 16 thread performance will be utilized in anything they use on a daily basis. The 4 core Ryzens may be a good i7 alternative for less ridiculous prices, making well performing 8thread CPUs available to people not made of money, but none of what AMD has shown so far seems to be really threatening intel flagships, so the most reasonable thing to do still seems to be buying a used last or second to last gen intel CPU from someone who paid the full retard price a year or two ago and now wants his extra 5% of performance. Now, if Ryzen forces everyone to actually start accomodating >8 threads, that will be something else entirely...

>>58356906
Not him but I want it to be near the same level so I don't get fucked into 20 lanes.

>>58356168
Here's a hint: Vega's HBM2 memory isn't being called that, it's being called a "High-Bandwidth Cache" and all memory-related operations on the GPU (and possibly from other places) goes through their HBCC (HBC Controller)

For all practical purposes, it functions as your average framebuffer pool, but the pieces are there for it to be also used as an off-die cache for a Zen-based APU.

Exciting times for us consumers.

>>58356168
>>58356960
Source http://www.anandtech.com/show/11002/the-amd-vega-gpu-architecture-teaser/3

>>58356960
>off-die cache for a Zen-based APU.
Is there any point to a Zen APU's graphics when you already got a 12tflop vega cunt in your PC?

>>58356422
>power advancements
lmao, see >>58355382

>massive iGPU increases
only matters for laptops and shit, everyone else has a dedicated gpu

>>58357012
>only matters for laptops and shit
So fucking what? That just means you aren't the target audience of the product, that doesn't mean no advancements have been made.

You dont CARE about the advancements being made is one thing, claiming there haven't been ANY advancements is just wrong.

File: 1438810686568.jpg (24KB, 499x499px) Image search: [Google]

24KB, 499x499px

>>58348218
how fast does it run gaems tho?

>>58356906
What part of "at the same number of cores and threads, AMD matches (and in edge cases, outdoes) Intel's offerings at the IPC front" do you not get??

This isn't AMD fighting Intel's IPC by going "MOAR COARS" this is, thread-per-thread, matching Intel's IPC with the same number of physical/logical cores; you don't match IPC at the same number of threads if you don't perform well per-thread, anon.

Are you being resistant on purpose, or do you genuinely not understand how this works?

>>58356559
Punctuation, nigger, do you know what it is?

>>58357079
>AMD matches Intel's offerings at the IPC front


LOLOLOLOLOLOLOL

>>58357091
>Intel babies literally chimping out with incoherent noises when presented with facts

>>58357128
Or maybe they're laughing at your DELUSIONS and as properly educated consumers know that shit you said is IMPOSSIBLE

File: BzZn4kg[1].jpg (118KB, 775x577px) Image search: [Google]

118KB, 775x577px

>>58355119

File: qt.jpg (28KB, 391x290px) Image search: [Google]

28KB, 391x290px

How is the IPC of a single core of Zen vs current offering on Intel in the same (expected) price range?

Not a gamer, but a programmer which like high

File: wahaha.jpg (367KB, 1269x1800px) Image search: [Google]

367KB, 1269x1800px

>>58357138
>properly educated consumers

Did you even glance at this thread? You guys are fucking idiots without peer.
I sure hope it's on purpose, it's kinda unnerving to share the same website with the braindead.

>>58355883
>those old times when transistors used to be 12mm and there wasnt justin bieber around

File: 1405187180638.gif (999KB, 280x210px) Image search: [Google]

999KB, 280x210px

>>58357138
>has literally been confirmed by third-party reviews
>DELUSIONS
>IMPOSSIBLE

>>58356994
The HBCC on Vega itself is for its own performance benefits and they're likely doing this to enable access to other forms of memory/storage for the GPU. Unlikely anyone will use memory on a discrete Video Card as a cache for the CPU.

But the fact that the HBCC exists in Vega means they can use that design for other places later, which is likely the case for making an APU with a Vega on it later.

HBCC looks like one of those "keystone" techs that's going to stay with us for quite awhile.

>>58357194
it matches it, why do you keep samefagging the same fucking question?

go to google for fuck's sake and stop grasping at straws

>>58357194
Zen is still behind Haswell-E per core with AVX2. Without, allegedly it could match Broadwell-E, but I highly doubt it given AMD's past propensity for overstating their capabilities.
Skylake-E is coming out in a few months, so wait for that.

>>58357194
Unless you're asking about it on a price basis, which we can't say anything about until we see some confirmed prices.

But on a IPC per-core/thread basis, it's matched.

>>58357245
Its just what they have on the Radeon SSGs.
A local storage cache for loading things faster. Doesn't have to interface with the CPU.

>>58357352
the SSGs don't have an HBCC and are based on Polaris, I believe, and don't have HBM or HBM2.

>>58357332
>AMD's past propensity for overstating their capabilities
Third-party reviewers aren't marketing slides.

>>58357398
Bulldozer?
Piledriver?
Fiji?
ReBrandeons?

>>58357398
Yeah because reviewers are TOTALLY not biased.

We see faggots on here all the time posting screenshots that cut off parts of benchmarks that aren't as flattering to their CPU, bias exists, even among your venerated "3rd party reviewers" you seem to fawn over.

>>58357382
That is how they operate, that controller already exists. The GPU is directly pulling from an SSD, it has a memory controller addressing memory that is not VRAM.
HBM itself is not a necessary part of this.

>>58357414
>>58357415
Show me a third-party reviewer that lied about the performance of ANY of those parts.

/g/ has, of late, been populated by the least informed crowed I've seen in years.

File: 1463207130232.gif (898KB, 540x400px) Image search: [Google]

898KB, 540x400px

>>58357497
>/g/ has, of late, been populated by the least informed crowed I've seen in years.
I'd have to agree, I mean believe it or not, we're getting people who actually think AMD has a viable CPU architecture, can you believe those morons? They must be just barely 18 years old.

>>58357511
So, this is what you resort to; shitposting?

where's those lying benchmarks, neighbor?

>>58357511
This, AMDtards are illiterate

>>58357552
nice samefagging

>>58357544
ask anyone who bought an FX series CPU based on early reviews how they fell about their purchase.

File: 2017-01-06 13_11_32-(0) _g_ - HAHAHAHAHAHA OH WOW Vega might still be questiona - Technology - 4chan.png (8KB, 404x170px) Image search: [Google]

8KB, 404x170px

>>58357558

>>58356407
take that shit to /ck/ you fucking moron

>>58357561
why don't you link to those benchmarks, then?

>>58357568
lol

>>58357568
grats, you know how to use paint.net

File: Screenshot from 2017-01-06 19-13-23.png (49KB, 1088x200px) Image search: [Google]

49KB, 1088x200px

>>58357568
How do I into HTML George?

>>58357595
Go ahead and try using paint in the ~30 seconds between the post and the screenshot being posted.

Your computer probably couldnt even load up paint in the time it took me to post that.

File: fuck you faggots.webm (304KB, 748x424px) Image search: [Google]

304KB, 748x424px

>>58357605
When you refresh the page the HTML resets, so here is a webM of it just to prove it.

>>58357646
can i have that css falaman?? I mean the general one not the (You) one.

>here is another you btw

>>58357605
fuck this noise, all the intel luddites are all awake and any hope of a engaging discussion regarding the finer technical aspects of various CPU-repated techs are gone.

Fanboisim is a fucking non-lethal plague upon humanity.

>>58357673
Just 4chanx with onee-san

>>58357646
nigga we know you used adobe premiere you can't fool us

>>58357680
>engaging discussion regarding the finer technical aspects of various CPU-repated techs are gone
lets have that discussion, all I have heard from AMDrones is about how it's gonna blow intel out of the water with zero proof to back it up.

File: 1447784761932.jpg (55KB, 640x480px) Image search: [Google]

55KB, 640x480px

>>58357712
you caught me

>>58357719
Ok, first lets hear your opinion on the L3 victim cache in Zen

>>58357753
I think it's a novel approach to the problem but am unsure how it will perform in actual daily use scenarios compared to intel's caching methods.

Also I am curious to see if intel continues with their L4 cache scheme that they've been using on and off since the 5th generation.

>>58357697
I thought 4chanx was discontinued as fuck.

>>58357781
Got updated just 12 hours ago.

>>58357780
Intel changed their internal caches with Skylake.
L4 is a expensive test which isn't working out all that well.

>>58357805
They''re still including it with the kabylake CPUs that have the Iris Pro iGPU.

Seems to provide very nice gains for iGPUs specifically, though I can envision some non-iGPU scenarios that could take advantage of a 64 or 128MB L4 cache.

>>58357826
I meant isn't working out all that well financially, it's rarely ever used, CoreM/15W ULVs won't use them because they're power constrained, 25W+ chips are already aproaching 2kg+ laptops and those usually have a dedicated GPU that shits on it for breakfast, even Apple isn't using them anymore.


I've only seen them in some decent quantity in SFF like the NUC or NUC-like things.


As for performance, while it does make some good bandwidth for iGPU it uses a lot of power and space, it's an off die piece of eDRAM that needs its own cooling solution so its at best, clumsy.

I'd love to see it in desktop though, Broadwell desktop with Crystalwell performed a bit better than Skylake at lower clocks.

>>58357891
>I'd love to see it in desktop though, Broadwell desktop with Crystalwell performed a bit better than Skylake at lower clocks.
This would be the only thing that would get me to upgrade from my 5820k.

If skylake-E or Kabylake-E has a 128 or 256MB L4 cache. And honestly even with that, I would be hard pressed to justify upgrading.

File: 1430439132870.png (88KB, 1082x1055px) Image search: [Google]

88KB, 1082x1055px

>>58355352
>you weren't called stupid enough as a kid

>>58357916
>If skylake-E or Kabylake-E has a 128 or 256MB L4 cache
These are GPUless chips, I really doubt Intel will waste so much die and packaging area to give a 10% boost at best to its HEDT line which is already sporting large die sizes, besides there's no bus for that L4 on the HEDT chip, since the bus goes on the GPU normally, which would just snip at their profits even more.

File: 1475650763600.jpg (31KB, 520x519px) Image search: [Google]

31KB, 520x519px

>>58348218
Duhmanon here. What am i looking at?

>>58357972
I would want it to be an on-die cache anyway, so it would require a re-design of the core as it is.

>>58357987
COOOOOOOOOOORES

>>58358003
That would be the most optimal, but integration is difficult and hella expensive.

>>58358024
Well with their ULTRA-high end server space finally getting those integrated FPGA's I expect to see intel attempt to put those FPGA's on-die eventually though I'd expect mid 2020s for this, and that's if something else isn't developed that supersedes this.

>>58358048
How's that integrated FPGA even gonna work? FPGA's are generally just huge masses of fast memory, where would that fit on die unless they plan to make the chip the size of your hand?

On package/offdie seems more plausible to me.

>>58358085
They're already at on package off die currently with FPGA integration from what I have seen.

Only place to go from there is on-die.

>>58358085
>How's that integrated FPGA even gonna work?
Same way anything else works. FPGAs are only large dies because they're made on old nodes, or they're enormous arrays for simulating really complex full chips. They're not trying to have the ability to simulate an ARM core in there or anything. Intel is pursuing the FPGA so they can have a programmable on the fly code path to make specific instructions faster.
They'll get them on die.

>>58358131
Yup, need encryption? Load up the encryption library and reboot.
Need 3D rendering? Load up the rendering library and reboot.
etc
etc.

>>58348218

I don't get why people are so focused on brands. Do we want technology to stagnate? Competition also drives prices down. I hate fanboys.

Get the best tech. Don't get your favorite sticker, you stupid faggots.

>>58357568
> Tries to make screenshot to prove he wasn't samefagging
> Forgot to actually edit out the (You)

>>58359247
are you retarded?
The post saying
>nice samefagging
should be the one that says (you) if I WERE both posters, I never denied I was the FIRST poster, I just proved I wasn't the 2nd one.


Learn to 4chan retard.

>>58348218
Holy shit! This is awesome!
It's the multicore power of a Xeon and the single core power of a i7!

>>58356020
>AMD only "rented" him to lay the base of the chip and then he went on to do something else.
he worked on it for 3 years out of 4 it was in development

>>58359911
He's an architect, his job is done when the ES gets fabbed, the grueling verilog debug work is done by others later.

File: Zen team.jpg (75KB, 715x573px) Image search: [Google]

75KB, 715x573px

>>58359911
>>58360169
I don't know where this meme came from.
Keller wasn't directly working on either Zen or K12. Jim Keller was the big daddy department head of all AMD's core arch development. He managed the teams working on the designs.
When Keller left AMD again the high level arch of Zen and K12 were finished. AMD announced in either May or March of 2015 that they had already started development on Zen+.
Pic related is the Zen team minus lower level technicians.

>>58355575
Phase count is NOT everything. 32 phases of 2 D-PAK (1push-1pull) per phase is still crap.

File: 1483674966232(UpRGB)(noise_scale)(Level1)(tta)(x4.000000).jpg (355KB, 1036x776px) Image search: [Google]

355KB, 1036x776px

>>58355119
I 4x upscaled it for you

>>58348218
Is it finally time for Intel to stop selling us 120mm2(1/3 are GPU) dies for $300?


Naaaah
Idiots will eat up anything.

>>58359741
AND YOU CAN OVERCLOCK IT EVEN HIGHER
PRICES WHEN GOD FUCKING DAMN IT I NEED A PRICE POINT NOW

>>58361470
Just wait. You don't want another Phenom-tier fuckup.

>>58361573
A phenom-inal fuckup is okay by me. Everything sounds too good, something bad is bound to happen.

>>58361599
> Everything sounds too good, something bad is bound to happen.
Probably Athlon64-tier fuckup aka not enough supply for demand.

>>58360217
She has zombie feet.
What
The
Fuck

>>58361651
That's what old age does. She's way over 40.

>>58361651
about normal for an (>)engineer

>>58361625

AMD's Jim Anderson, SrVP and GM, said that they will not do a paper launch with Ryzen. They will launch the whole line up at once with stock on day 1. They say they want an early to mid Q1 launch. Indications are they will not release in January, so that leaves either sometime in February or early March. So Ryzen is in all likely hood entering production right now if they want to have enough stock for everyone.

> http://www.pcworld.com/article/3155129/components-processors/amd-says-its-zen-cpu-architecture-is-expected-to-last-four-years.html

File: linus steals kellers booty.jpg (184KB, 900x1200px) Image search: [Google]

184KB, 900x1200px

>>58362176
So, no fuckups for now? Now we need prices and soldered heatspreader and i'm sold.

>>58362239
The guy who writes for bitsandchips said the IHS is soldered.
Godavari and Bristol Ridge are also soldered for what its worth.

>>58362303
>8 cores
>3.6/4.0 clockspeed
>soldered IHS means actual fucking OC potential
AMD, please, for once since Athlon days, do not fuck up.

>>58362239

> No fuckups

Look man I like AMD, but it IS AMD we are talking about. They WILL fuck something up. Just hope it's marketing, branding, or something stupid like that.

>>58362346
B-but it seems like good old Athlon days when Intel went on LITERAL suicide watch.

>>58362340
Word from everyone in the know is that later steppings of Ryzen are good to go. No known bugs, everything ironed out.

>>58362346
Its a guarantee that the marketing is going to be fucking stupid. Remember the cyborg video they made for their APUs?
They made a whole short story, even had a comic book to go along with it, about a robot police task force called the Cores or something like that.
https://www.youtube.com/watch?v=MQcjEA3it90

How much money do you think they wasted on this? It was a whole little series.

>>58362430
>How much money do you think they wasted on this? It was a whole little series.
Didn't they gas their marketing division at least once since then? Now i need prices.

>>58355886
dont get too excited, nobody actually bit your bait

>>58355588
You know that voltage setting you adjust when trying to get a higher overclock? The one you should be able to dial in down to 0.01 volt? And all the raw current flowing into the chip that needs to be squeeky clean DC with virtually no ripple or else leakage within the CPU itself starts causing all kinds of crazy shit potentially even damage?

That stuff is controlled by the board's VRM. Higher quality modules maintain a given voltage better, higher capacity caps with low esr filter noisey voltage better, and more phases distribute the workload amongst clusters of these components.

It's actually so important that often times the only real difference between an "ok" motherboard for overclocking and an "excellent" motherboard for overclocking is VRM design and an extra bios in case you fuck up bad enough you cant even get the mobo to post.

>>58362430
>>58362448

I would completely axe the marketing division except for maybe a couple guys who understand the tech side a bit more to explain shit to outside contractors. Hire actually good outside marketing companies. It would work out so much better for them to have people who actually do the job everyday in their respective field of expertise and area. We would hopefully not see shit like the robot and this whole "JOIN THE REBELLION MOAR DRUMZ" kiddie shit.

File: Dr Lisa Su.jpg (80KB, 650x450px) Image search: [Google]

80KB, 650x450px

>>58362448
They've laid off a bunch of people over the years. People jumping ship, some people getting cut in the financial restructuring, others fired for failing to meet goals.
However Lisa Su herself shitcanned a select handful when she became CEO.

Colette LaForce came to AMD from Dell where she was CMO and Global VP of Services. At AMD she took the role of CMO and SVP in 2012.
Su personally told her to get the fuck out. Allegedly of course.

The Fiji press event, appealing to venues like E3, that was organized by Su. The lady is a badass.

>>58362788
>chinkmama makes AMD, quite literally, great again
>everyone calls AMD poointheloo company
Why?

>>58362852
Head of the GPU department is an Indian guy, and memeing kids take anything they can grasp and run with it.

>>58362864
That "Indian guy" is one of the most famous and respected GPU architects on the planet and has been for the last decade.

>>58362864
>Raja Koduri
>literally papa radeon
I still can't grasp why.

>>58348240
This. Also don't forget the possible OC up to hopefully 5GHz.

>>58359465
That makes no sense. You're saying you would be accusing yourself of samefagging?

>>58355589
You're actually retarded. Do some research on Ryzen before making you keep making yourself look bad.

>>58355033
Ryzen IPC matches Broadwell-E.

>>58355382
>Muh power consumption

My i7-6700 is at 4GHz stock...
The new AMD products are too slow even at 8 cores to compete.

>>58362871
Then why was he staring at a twitter feed during the 480 reveal while he lied and told everyone that 2x480 were better than a 1080 and they would cost $199?

Tell ya why, because hes a poojeet.

>>58356076
Dude cancel it, it's not worth it

File: why.jpg (30KB, 288x499px) Image search: [Google]

30KB, 288x499px

>>58355382
>O/C to xG