AMD are rumoured to be creating a consumer-oriented 16-core Ryzen CPU

>add more cores n shiet
>2017
>not focusing on IPC for the consumer market

MORE CORES
O
R
E

C
O
R
E
S

now THIS would be faildozer all over again

File: 1484660055551.jpg (124KB, 720x719px) Image search: [Google]

124KB, 720x719px

>>59463995
>>59464002

Not only more cores but it 8c in one of its 2 CCX. Power consumption would be amazing.

>>59463981
would love to have one, but even the 1800x is way beyond what i can afford for a cpu

in the 90s i dreamed about what the future would bring, little i knew it was dual core pentiums. i thought it would be 20 core cpus with 1TB of ram

>>59464068
R5 and R3, breh.

>>59464068
I think the the 1800X is over kill for my everyday use. I'm looking between a 1600X and a 1700

File: 1484915884086.png (376KB, 651x481px) Image search: [Google]

376KB, 651x481px

I hope they push core count even more
Intel will be forced to do it too, that would be huge win

>>59464133
that's not how it works. they would have to get applications to use more threads, otherwise intel wont change shit.

File: 1489331451216.jpg (20KB, 504x415px) Image search: [Google]

20KB, 504x415px

>>59463981
>moar cores = faster cpu

Heavy as fuck video codecs like Daala might have a chance if future CPUs have 16 cores or more.

will naples be on lga or pga?
I was wondering what would the x399 be called, if simply that or am4+

>>59464105
I'm going 1700 because you can overclock it (not rven going to) but you can't add cores to a 1600.

I hope they test it on 720p games

>>59463981
>This should be competing with the Xeon and of course 6950X Intel offers for $1700~$1800USD,


BUT
WHERE
ARE

M
Y
GAMING
BENCHMARKS

>>59464178
It wont be AM4+ cus that's consumer name and they wont make another consumer socket until DDR5 arrives.
Probably PGA since Ryzen is PGA, unless the pin count is too high?

>>59464248
Uh nvm the OP says it's LGA.

>>59464198
Yeah, the 1700 is the sweetspot imho for people who need a lot of power to get the job done. It's 4 years that I do everything on a laptop that costed me about 250€, with a ridiculous i3-3110m, 4gb of ram and a 5400rpm hd, and I'm simply dreaming about a hardcore build, which ryzen makes possible, for developing java, machine learning, music production and a lot of other stuff that I find hard to do because right now I'm bottlenecked. Thanks based ryzen.

File: 1483626643500.jpg (66KB, 568x612px) Image search: [Google]

66KB, 568x612px

>moar corez yet again
>2GHz

>>59464429
>MAKE LESS CORES
Intel

>>59464068
One of the computers where I work has a 22 core Xeon and 1tb ram

File: 1464804859737.jpg (52KB, 568x612px) Image search: [Google]

52KB, 568x612px

>>59464439
But deliver.

File: 1467296306101.png (91KB, 653x726px) Image search: [Google]

91KB, 653x726px

A-all Issues have been totally ironed out.

AYYMD HOUSEFIRES

File: tmp_10364-images(3)2043140089.jpg (62KB, 554x554px) Image search: [Google]

62KB, 554x554px

72 x86 cores

File: 1488404508616.png (316KB, 882x758px) Image search: [Google]

316KB, 882x758px

Now my 8 core risens is obsolete..

File: 1449999756930.jpg (49KB, 423x389px) Image search: [Google]

49KB, 423x389px

Remember to ignore and report literal intel shills!

File: letdown.png (609KB, 854x640px) Image search: [Google]

609KB, 854x640px

And beaten by a fucking i3.

File: amd.png (1MB, 1228x1502px) Image search: [Google]

1MB, 1228x1502px

W-wait I can't afford that now..

>>59463995
Encouraging shitty programming does not result in a better end user experience.

File: 1488919452885.jpg (514KB, 960x686px) Image search: [Google]

514KB, 960x686px

And the importance of FPS goes even more downhill.
>just wait

>>59463981
if it's within 5% of the r7 on single threaded and double multi threaded I'll buy it

>>59464439
>FAKE MORE COARZ
t. intel

File: 1483629357673.png (181KB, 293x397px) Image search: [Google]

181KB, 293x397px

And with Intel's grip on the whole CPU market we are totally going to see proper core AND CCX utilization. Yeah.

>>59464597
>FPS
>>>/v/
this isnt a fucking gaymur chip

>>59464692
I thought I buried you underground for good this time, Oldman.

>>59464721
Not yet, newfag.

>>59464768
The worms will be enough this time around.

>>59464503
>P54C cores

toppest fucking kek

>>59464680
There will be a respin of the AM4 lineup to fix the DF issue, and this new HEDT lineup won't have the same issues as the current desktop parts.

>>59464692
Ryzen 7 literally said "for gamers".

I'll love when the v4 xeons arrive at ebay for 200 usd each.

>>59464859
...on a footnote. Try harder, shill.

>>59464859
>Unironically believing AMD propaganda
>Believing any kind of propaganda

Why didn't you wait 'till the reviews came out to see how well it did outside of gaming, anon?

>>59464855
When?
>just wait

>>59464889
Why do you care? You can buy Intel now.

>>59464921
But muh futureproof, obviously ;)

>>59464889
t. impatient manbabby

>>59463981
How do they plan to do that?
Current motherboards have trouble not catching fire with a 8 core 95 watt TDP CPU.

>>59464939
H2'17. They will have different model numbers, of course.

>>59464947
>just wait
Okay, grandpa. I thought you had business with maggots in the ground.

>>59464225
Nah bro, 480p is the new standard

So all trolling aside, I do play gaymz but also do some editing and rendering.

What are the odds of this actually being on par with the 6950x?

>>59464987
>no 240p

pleb

>>59464968
Post link.

>>59465003
is(.)gd/WZMFMM

>>59464988
It would be similar with 16 core Xeon.

>>59464998
>240p
>not 144p

>>59465027
;]

>>59463981
That definitely is not a public knowledge, just a wild rumor that hasn't appear in any halfway reliable site.

So pretty much don't believe this (for now).
I'll believe it when BenchLife.info confirms it.

t. cretin working as hw journalist

>>59464178
It'll be on BGA. This has been confirmed.

So will the 16 core Snowy Owl.

>>59464048
If it appeared, it would be 16core made of two Zeppelin chips, so 4xCCX.

>>59464259
OP is wrong.

It's a rumor from a fucking forum from a guy that talks out of his ass.

Snowy Owl might be brought to HEDT as a X299 competitor at 140w or 180w TDP instead of <100w TDP for the server part, but so much of that other shit the dude says is blatantly wrong.

Looks like a server oriented CPU. The average person probably wouldn't buy it because it's completely shit price/performance but it might be more affordable than similar CPUs in the server market.

>>59464503
>72 x86 cores
18 cores with 4-way SMT, you mean.
And the biggest lie marketing campaign.

Fucking joke.

>>59464811
The current ones (you said 72 cores, so no excuses) are Silvermont + custom AVX512 unit and 4way SMT. Not bad actually, but not for PC of course. Would be fucking cool to have it for desktop though with 96 GBs of DDR4.

>>59464859
No, it said for enthusiasts and prosumers.

>>59465103
Dude, you know shit. It is 72 actual cores. 288 threads. Next time get your information before shitposting.

File: intel-knl-dieshot.jpg (921KB, 2000x1319px) Image search: [Google]

921KB, 2000x1319px

>>59465103

Sure it is. Intel just disabled 90% of the cores on the die and then lied about it.

>>59465288

>This time, for reals we'll build a gpu that works!

>>59465288
Oh, cool.
Still performs shit compared to a GPU.

Looks pretty, tho, I'll give it that.

>>59465357
It's a standalone CPU with up to 384 GB of RAM and 16GB high bandwidth cache, so it's much better than GPUs in many ways.

I like the idea even as AMD fan. With they did it.

>>59465375
Doing dual socket with an ARM cpu in one and x86-64 in the other seems like it'd make more sense.
Or an iGPU so you can dual socket them, unless that's what it does?
My understanding is that it's all x86-64 cores which holds it back.

>>59463981
Terry might like this

File: Make_more_cores.jpg (197KB, 800x1000px) Image search: [Google]

197KB, 800x1000px

>>59463981
>So Cinebench, and all other related productivity programs.

Lol Cinebench isn't a productivity program.


And speaking of those... Raisin doesn't stack up so hot to Intel's 4 core chips.

This is Bulldozer all over again! Lol my sides, I can't even...


MOAR RAISIN FLAVORED CORNZ!!!!!

>>59465516
Are you SiliconDoc? Your tactics seem to be as annoying as his so I wouldn't be surprised.

File: 1463606423970.gif (152KB, 400x267px) Image search: [Google]

152KB, 400x267px

Better parallelism software when?

>>59464563
shitty programming and human error and stupidity just piles up when you have more complicated configurations such as 16 thread processors, you need to do more to get proper use of such a processor and in practice it will be shit every single time

>>59466517
It's good that intel isin't coming out with a 32 core design then. You obviously know better than them, right?

>>59466545
damn didn't know they have this consumer oriented 32 core processor coming

>>59464068
What's your job? Just curious, not gonna bully

>>59466553
Intel wants you to pay $10000 for that privelege.

File: 1484442295292.png (204KB, 600x528px) Image search: [Google]

204KB, 600x528px

>>59464855
Just like anyother drone, full of shit.

>>59463981
I DON'T GET WHY PEOPLE STILL POST ON THOSE THREADS

THE ONLY NON-BULLSHIT THREADS ARE THE CONSUMERISM ONES AND THAT'S SAD

>>59466655
almost as if it's not for regular joe consumers

>>59463981
Hopefully they're dual socket or I'll have to go for the opterons where I can expect Supermicro to rip me off on prices

>>59466491
Much in gaymes aren't parallizable.

h265 doesn't seem very parallizable, either.

For games, you have to design gameplay around it but most designers aren't programmers.
One example is that Civ VII could have AI do turns in parallel. Some turn based games do do parallel turns. But they aren't the same, they require different rules designed to work in parallel.
Currently turns go in an order which requires the next turn to know what the previous one did. Not much can be parallel-ized there.

And for other software, often they rely on shared memory. Parallel-ization works well when you can "tile" something into sections.

File: IMG_1553.jpg (546KB, 1972x1376px) Image search: [Google]

546KB, 1972x1376px

>>59464597

https://twitter.com/CPCHardware/status/843105726298054656


Kek

>>59466992
32c@3GHz with 8 DDR4 channels and 64/128 PCIe lanes in 180W seems pretty to me honestly.

>>59467309
3hz 32core would not be 180W TDP. It would be 250-260.

>>59467309
I actually expected somewhere around 2.7-2.9 under 200W, wasn't so off the mark.

>>59466992

Intel is going to be sweating like a peado in a playground at the prospect of server zen - its the one market where the real money is and they can't risk AMD gaining traction.

>>59466992
>B0 stepping

Oh this is gonna be good.

>>59466992
THIRTY TWO FUCKING DIMM SLOTS ON A 2P MOTHERBOARD


SHIT UNHOLY FUCK THAT'S FOUR TERABYTES OF MEMORY USING 128GB RDIMMS!

FOUR
TERABYTES
OF MEMORY
ON A DUAL SOCKET BOARD

>>59467613

...but can it run crysis?

>>59467642
No, Windows doesn't support that much memory.

>>59467658

Stallman be praised. Saving productivity from the evils of gaming.

>>59467309
It's 128PCie lanes with a single socket, and 128PCie lanes with two sockets.

Either way you get 128PCie lanes, which is absurd amount of I/O.

I don't think motherboards actually have enough SATA/SAS/M.2/U.2 slots to populate that much bandwidth even with 4 GPUs running their course.

I guess you could do some riser card fuckery to get the most out of it.

>>59467762
You're forgetting that the dies used on Zen (and thus Naples) are SOCs and that the Sata drives will likely link directly to the CPU.

These processors will not be wanting at all for I/O

>>59467613
This is one of the most interesting parts.

While the $9000 Xeon CPUs do support slightly more memory, they require (presumably) much more expensive 256GB DIMMs.
And while the $5000 Xeon CPUs support almost as much, over 3TB, they also require 256GB DIMMs to near their memory limits.

Actually... do they even make 256GB DDR4 DIMMs? I can't find them.

I wish Snowy Owl really would come as HEDT 140w/180w TDP parts with all the usual consumer chipset features, and dual socket motherboards.
I'd love to put 2 12 cores with the same 3.6/4.1Ghz clocks on a board.

AMD could open up a whole new market for $2500+ PCs that are actually sort of worth it. Like the good old days of dual Opterons.

>>59467805
This is nice point too, these motherboards won't have chipsets and will have lower platform power on top of lower package power.

I don't think Skylake-EP/EX is a SoC.

>>59467762
You already see some servers from Supermicro etc. with 10x 2.5 u.2 drives in the front of a 1U box, so a 2U with 20 drives would use 80 PCIe lanes and have 48 left for some absurdly fat network IO.

6-7x GPUs plus networking could end up common in 4U boxes as well.

>>59467613
>4TB

How much memory does Zen support? 6TB per socket?
Or 6TB per zeppelin die?

If someone comes out with 256GB RDIMMs in the near future I doubt they'll be using Intel's chips in this case.

>>59467816
128GB is the largest one so far, and it came out in late 2015.
So I'm expecting 256GB or 192GB DIMMs in the next year or two.

Just a reminder one of these sticks cost more than your life.

>>59467853
Much more likely its 6TB a socket. 6TB per die would likely require a shit ton more memory channels on the die to spread the strain out.

>>59467853
Nobody knows what capacity RDIMMs and LRDIMMs Naples will support.

DDR4 in theory supports up to 512GB modules, but fuck if things don't start getting priced silly at >32GB.

>>59467895
Just a reminder that folks running oil and gas simulations don't care how much hardware costs and just buy whatever the fuck they need, and they usually need more than what the market can give them.

>>59467895
Yeah. $850 for 64Gb stick.
$1950 for 128Gb.
32Gb, meanwhile, is "only" around $250.

Can guess that 256Gb would be around $4500-$5000.

Really the bigger deal here is that you could use those 32 DIMM slots, compared to the 8-16 of Intel, to get 2TB of memory using 64Gb sticks for about 2 grand of savings compared to using half as many 128Gb. Or 1TB with 32Gb would only be more huge savings if 1TB is enough.

Compareable 1TB and 2TB memory dual Naples servers are going to be ridiculously cheaper than Intel even if AMD prices Naples at a compareable $4000-$5000. And the CPUs probably cost AMD less than $500 to make.

I'd think they'll absolutely destroy if the software support is there, and especially AMD-V is working good enough.

>>59467960
Well for those people, can you even make a 4TB Intel system?
They only have 16DIMMs max, right? 16*128 is only 2TB even though the CPUs support more.
I googled and did find some that support 32DIMMs. I don't think it's in a configuration that utilizes it well. So I guess they can.

I'm not sure Ryzen would be ideal for them, though. Its float math perf still seems inferior to anything Intel has to offer, and that's what they have Vega Instinct for.

>>59468091
There's no way a 4 channel chip can run 32 DIMMs natively, it's either a 4 socket board or its running the memory in half speed.

>>59466898
It would be good for 4k or 8k processing with appropriate settings. H.264 lacks this kind of intra threading iirc.

https://www.parabolaresearch.com/blog/2013-12-01-hevc-wavefront-animation.html

File: desktop phenom logo.jpg (159KB, 894x588px) Image search: [Google]

159KB, 894x588px

Blah, nothing consumers do require a 16 core chip, least not yet. Considering that still a lot of programs are single thread so clock speed is king not to mention how good said chip is at single thread tasks. Let's see most demanding tasks most everyday people do is check e-mail,surf web, or watch youtube vids. Oh and dump a shit load of pics from there point and shoot cams to the hdd for safe keeping. Really, how is a 16 core, much less anything over 2 cores help them any? Now step it up, to your power user/gamer, 4 core will help them, provided the game or app really uses those extra cores. More ram/Higher clock speed is the real way to gain more performance boost. that and a ssd as your boot drive. (btw, ye old little factoid = Windows 7 pro/ult x64 can handle 192GB of ram,Win Server 2003 sp1 x64=1TB,Win server 2008r2 ent =2TB,WHS 2011 = 8GB)

Wow get a load of the intel crybaby cucks. Enjoy your overpriced, overspecced jew spyware.

retarded rumomongering. there is a sixteen core cpu coming out, it's a server part with 4 ccx's.

File: AMD-TOLD.png (16KB, 441x495px) Image search: [Google]

16KB, 441x495px

AMD's back baby!

>>59469256
>AMD's back baby!
Back into the trash.
>>59468439

>>59468390
>overspecced


the horror!

>>59469309

Whatever pussy, some of us get paid to do real things with computers, not pay to play badly coded games. Enjoy your overpriced benchmark epeen.

>>59469372
Yeah, get paid to shill their product, fellow RedTeam+ member.

>>59469395

Oh man the irony of that comment is off the scale.

>>59463995

>muh ipc

Enjoy playing your artificial benchmarks kid.

File: 1487838437461.jpg (18KB, 267x297px) Image search: [Google]

18KB, 267x297px

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!

I'm not ready yet !
STOP, I just need a bit more ..uggh.. Time.
*sweeting* Just buy the 7700k and the 6950k in the mean time its better than anything amd has.
Please don't buy ryzen you can see its trash you don't need I-I-It believe the Gaming bench marks no need to worry your self about non gaming right?
Right?!!
ryzen is a flop.
Don't even dare think about that poor fag company Y-you're not going to leave me right anon? ... I-Its going to be okay

Shit ,,,,shit ..shit
I- I have to think.
I'M N-NOT S-S-Scared of you I swear.
J-Just a little more time is all.

We'll crush you flity peasant AMD its not over yet.

>>59469256

A lof of the toddlers on here wouldn't even have been born when that was a meme.

>>59463995
because their ipc is already on par with intel

>>59463981
12-16c, 4 DDR4 channels, and 64 PCIe lanes?
I'll probably end up still wishing I had waited for Zen+/Zen2/Pinnacle Ridge/whatever, but this could end up ridiculously tempting.

The competing Xeon Broadwells models are only:
> 16c - 2.1GHz/120W/$1700, 2.6GHz/145W/$2900
> 14c - 1.8GHz/75W/$1500, 2.0GHz/105W/$1400, 2.3GHz/105W/$1800, 2.4GHz/120W/$1700, 2.6GHz/135W/$2100
> 12c - 1.9GHz/75W/$1400, 2.2GHz/105W/$1200, 3.0GHz/160W/$2100

A 16c/3GHz/130W/<$1500 part would be easily achievable and would throw Xeon sales into complete disarray.

>>59470030

Whatever, AMD are pretty good about socket compatibility so you won't have to throw out your mobo if they drop a compelling upgrade.

File: some-people-just-want-to-watch-the-world-burn-humor-images-48.jpg (67KB, 620x803px) Image search: [Google]

67KB, 620x803px

Hi!
I'm buying a 1800x for gaming. (:

>>59470108
Nah, it's more that Zen1 is good enough for me to consider building a Ryzen setup, but I feel like a lot of this shit ironed out in Zen2 will be focused on server/MCM improvements, so spending $1k on a CPU might not be the best idea if I don't truly need it.

>>59470242
NOOOOOOOOO PLEASE BUY INTEL, MASTER GOLDBERGSTEIN WILL BEAT ME MORE AND FORCE ME TO POO IN THE LOO IF I DONT SELL ENOUGH I7 7700K's

>>59465516
You didn't know Cinebench is based on real-world rendering software by Maxon?
>these /v/ (man)children

>>59463981
16/32 is sexy as fuck with zen ipc, would buy

>>59468148
LGA 2011 (server) supports 12 dimms, so will LGA 3647 (it is six-channel, but only 2dimms/channel).

So Naples supports 33% bigger capacity with the same modules than Xeons per socket, because it's 16DIMM/socket.

>>59470502
16 cores? why do you need 12 cores? is 8 cores really necessary? i think 4 is a bit too much? 2 cores should be enough goyim

>>59466898
H265 is well paralelisable, there is WPP and frame threads to use. If it comes to worst, you can reduce the max block size to 32x32 from 64x64 -> 2x more threads from WPP.

I'm pretty sure you will saturate 16thread CPU with 1080p encode.

>>59466731
Dual socket always means Opterons, making dual platform for consumers doesn't make sense IMHO.

It would be more efficient to just release the octochannel/32cores.

What is better for single thread/core operations

>>59470610
intel

>>59470610
>>59470635

Since thew 3rd of march 2017 every piece of x86 software has become single threaded to the extreme.

>>59470610
Amd

>>59465058
Its common knowledge, amd has said over a year ago they were making a 16/32 core thread cpu.

The current thought is that instead of server hardware and server platform exclusively, it will be amds answer to intels high end cpus for consumer workstations.

>>59470610
amd and intel are equal in the regard, its the multithreaded shit that there is a difference in and it shows intel getting btfo hard

>>59463981
>muh cores xD
AyyMD really needs to fucking go bankrupt.

>>59470396
My plan is buy a good motherboard, possibly decent ram, or possibly burner ram till the ram issues are sorted, and upgrade to zen + or zen ++ when they come out, and put my old 8 core into a crappy but decent motherboard to offload render loads to, possibly an overpowered emulation box for the living room to play games with family.

>>59463981
>The gaming issues that were causing the Ryzen AM4 CPUs to behave erratically to say the least have been ironed out
How would it solve anything? It's still the same Zen architecture. It's still a bunch of CCXs linked with the infinity fabric. There are still plenty of games out there that will rely on higher clocks and IPC vs more cores.

I mean, the quad-channel memory controller and 16 cores/32T would be a great for a workstation, but I see fucking nothing that would makes this idiot claim it would be better at gaming. Especially since there's no way it'll be clocked at 4Ghz.

>>59470731
Yes, muhcores and muhpowerefficiency are the only things that matter in bigdick CPUs. Corelets pls go.

>>59463995
But IPC is good now.

The only thing left to do is add more cores. Th

>>59470751
Because it's not intended for gaymen you sumbfuck. It's lower-end E5 competitor.

>>59470731
kool post shlomo

File: 1489882762383.jpg (104KB, 803x688px) Image search: [Google]

104KB, 803x688px

Intel i7-6950x

>>59470610
https://www.cpubenchmark.net/singleThread.html

7700k

>>59470706
>amd and intel are equal in the regard, its the multithreaded shit that there is a difference in and it shows intel getting btfo hard


lol in what way? intels 8c16t shit is still dominate over walmart brand amd

one new release at a cheaper pricepoint and its amd who?

for christs sake intels 4c8t is still better on average than the best amd can stroke out

>cinebench mullticore doesn't count
>cinebench single thread doesn't count
>blender doesn't count
>power effeciency doesn't count
>linux kernel compile time doesn't count
IPC doesn't count
>only ARMA 3 at 720p counts
>only raw clockspeed counts

Looks like a xeon-D competitor

>>59470767
Catching up is not enough

>$shitgame only runs at 150fps on ryzen deleted

>>59470995
This. Literally no point to buying Ryzen if it can't even beat Kaby Lake.

>as good as broadwell at half the price and half the power consumption
>buh buh IPC
>kaby lake don't scale cores
>buh muh single thread

File: 1424715856582.jpg (32KB, 312x342px) Image search: [Google]

32KB, 312x342px

>>59471067

>sandy bridge is all you need
>shit, AMD is only a hair behind kaby lake
>ahaha, kaby lake is the MINIMUM you need, anything else might as well be talking to babbage.

>>59470894
>cpubnechmark
go back

20fps less at 320x240 res,literally unplayable

>more memefinity fabric bottlenecks

>>59470977
>linux kernel compile time doesn't count
most impressive thing i saw out of ryzen benchmarks desu

thing is a beast at compile jobs and will make a lot of devs and gentoo users very happy

>>59471464

I'm sorry but those don't count. Ryzen is a gaming cpu and you can't PLAY kernel compile time.

>>59471235
you speak like you dont understand the point of testing at that resultion. for a tech board many people are utterly clueless

So, we should...
Just Wait™?

>>59471575

Pick your poison

>640x480 totally shows off what a cou can do when gpu is a non factor. 120fps vs 150fps is world changing

>640x480 is borderline synthetic as nobody plays at that, 1080p or high ultra is a real world scenario and 70fps vs 80fps is irrelevant because adaptiive sync exists

https://www.youtube.com/watch?v=4HERHZLFnqE

>intel
what a joke

>>59464884
I'm just saying that everyone who says that it isn't a gaming cpu to defend benches are not reading correctly

>>59471982
do you have that video where cooper is the intel ceo, and there's a song playing for some time with that woman holding the ryzen chip?

>>59473327
never seen it, link it if you can find it

>>59473376
I doubt I can find it, since i've looked for it for some time now

>>59463981
Intel are already hurting in the price/performance market, the $500 1800x out performs Intel's 6900x and it's $1050.
Thank fuck for market competition, Spintel have been miking everyone for a good decade or so. Their cucks will line up get milked every year, nothing too new will change for the fanbois.

>JUST ADD MORE CORES

>>59466972
>not reading the image
I don't even fucking understand how someone can post this image in good concious

>>59463981
so another cpu with no use in the consumer marketplace?

nice, i bet it will be at least $1500.

>>59473732
i bet you have autism

File: wake me up inside.png (69KB, 256x256px) Image search: [Google]

69KB, 256x256px

>>59465103
>>59465288
You clearly have no significant familiarity with this processor, nevermind direct experience.

>>59465113
You don't actually get access to 72 cores, though. You get access to 68, with 4 reserve as management cores.
But you don't always get 4 management cores either, because lithography is hard and yields are a harsh mistress. You get UP TO 4 management cores, and knowing how many you want/need versus how many you get... it's a big mess. And of course Intel isn't super transparent up-front about this.

But back to the point: 68 cores. Cool. That doesn't divide into powers of 2 for shit, and the NUMA domains are wonky as fuck.

That is not to say that I'm not impressed by the Xeon Phi. The distance they came from the original Larrabee shithow, to KNC, to KNL is impressive, and the machines you end up with can do some amazing things when paired for the right workload. So while it has yet to fully mature as a platform for most HPC workloads, I do expect it to eventually become something great. Just a big uphill battle to make use of it right NOW, where you have to take on all the risks of an early adopter.

TL;DR:
Impressive? Yes.
Mature? No.
Would I love to have one next to my desk to play around with? Yes.
Would I want to be forced to use KNL (in current form) when my job might be on the line? No.

>>59464403
just build home server for yourself with phenoms from ali for computing

>>59473767
yes i have autism because i know the average consumer has no use for a 16 core processor.

you caught me.

8/16 2GHz at 35W
My body is ready

Any indication on how well the 6 core Ryzen compete with unlocked i5? I'm never spending more money than that on a CPU.

>>59474153
1600x=1800x
There is only one ZEN core.

File: [HorribleSubs] Amaama to Inazuma - 03 [720p].mkv_snapshot_01.40_[2016.07.19_10.42.36].jpg (157KB, 1280x720px) Image search: [Google]

157KB, 1280x720px

>tfw waiting for my $70 4c/8t athlon
wew

>>59471645
Point is that it demonstrates maximum potential difference which WILL be an issue 5 years out when your GPU is no longer bottle necking everything

>>59474188
it doesn't.

https://www.youtube.com/watch?v=vA0vAYCiDds&t

here is two xeons that almost didn't work for games when they came out.

so the latency/core-parking/scheduling/SMT-issues remain unfixed?

ANOTHER
MASSIVE
DISAPPOINTMENT

>>59474749
ANOTHER
MASSIVE
ASPIE

>>59471067

>1/3 the price
>no reason

>>59474782
I like my CPUs working perfectly in all major operating systems without me going through 2,145,632 configurations and tweaks, tyvm.

>>59466517
>shitty programming and human error and stupidity just piles up when you have more complicated configurations such as 16 thread processors
There really is no excuse for the programming ethos of the early 2000's to be hanging around.

Fucking everyone has at least a quad-core these days. I don't give a shit if game designers have to treat two physical cores as a single thread, get the thing done.

>>59474804
because hes a autistic shill, ignore it

>>59473732
>so another cpu with no use in the consumer marketplace?
>nice, i bet it will be at least $1500.
It would be great for 3d rendering. 3d rendering on CPUs is very strange at the moment. 3d rendering on GPUs has become popular however that is largely because nvidia will sell you 1200W of GPUs for less than $3000 and you can fit them in a single PC case.

Intel only sells server chips with 130w for $8000. If you're not concerned with power use and just want to cram processing power into an atx case CPU manufacturers offer nothing suitable. If AMD could offer a 300W single socket house fire product with 16 and 32 core chips clocked at 3.6ghz+ it would be popular for workstations. Likewise a 2p 600W version would be popular. All that is needed are motherboards with appropriate VRMs and AMD to sell unlocked or high frequency server chips.

File: 4bc7449676b11131cd5c6eb84700b601.jpg (171KB, 1080x1710px) Image search: [Google]

171KB, 1080x1710px

>AMD's 500$ CPU performs worse than Intels 350$ CPU

>RX 480 is a total disaster and got hyped for months to compete with high end Nvidia cards but can barely compete with the GTX 1060 and loses in almost every case

>Vega rumored to be rehashed Polaris

You can cry, you can shitpost, you can meme, but none of that can change the real outcome, there is a reason why Intel and Nvidia are dominating the market for years, I don't want to defend their consumer practices but AMD showed that they are not good enough to compete.


I wish for a total new CPU or GPS manufacturer but due to complex of the materia the chance is near 0.

And here my favorite Indian bait picture.

>>59475693
>Vega rumored to be rehashed Polaris
The fucking what?

>>59475138

Yeah that would be pretty sweet. 20 years ago we called those desksides and they cost a bomb and it's cool that consumer gear is surpassing it.

>>59468227
I thought H.264 have threaded encoders and decoders that work fine.

I thought it's H.265 that does not.

The 8-core CPUs are already too expensive for most consumers, how the hell are they supposed to release a 16-core """consumer"""" CPU?

>>59463981
>And it is insanely quick in the tests that Ryzen is already excelling at. So Cinebench, and all other related productivity programs.

OP proves AMD made a CPU that the only goal is to make good cinebench and 7zip results

>Basically a benchmark CPU

>>59477436
Prosumer market.

>>59468148
>>59470552
Ty.

That's what AMD meant about leaping ahead instead of the incremental stuff Intel has been doing, right.

>>59475693
>Vega rumored to be rehashed Polaris

You're a moron. There's no such rumor except the one you just made up, which is easily proven false.
Kill yourself.

>>59467805
>These processors will not be wanting at all for I/O

typo? I dont understand. Is this a good thing or bad thing for Zen IO?

>>59478679
The processor waiting for something else is in 99% not a good thing.
I/O included

File: edgy.png (220KB, 373x327px) Image search: [Google]

220KB, 373x327px

>>59474782
>anyone who doesn't want what i want is autistic.

>>59464597
underrated screen tearing