He's a big boy

>>60710638
>For you

File: 1496380711245.jpg (270KB, 1919x1080px) Image search: [Google]

270KB, 1919x1080px

>>60710638
You call that a CPU?

File: CPOO.jpg (190KB, 1919x1080px) Image search: [Google]

190KB, 1919x1080px

>>60710660
>t. 5'5" cpulet

File: 1496342394694.jpg (200KB, 1919x1080px) Image search: [Google]

200KB, 1919x1080px

>>60710638
Healthy at any size. Don't judge me by my pin count. I get it from my family.

File: depositphotos_8321419-stock-photo-construction-trowel.jpg (38KB, 1023x796px) Image search: [Google]

38KB, 1023x796px

>>60710638
Lets apply thermal paste

File: 4u_rackmount_server_case.jpg (341KB, 1024x683px) Image search: [Google]

341KB, 1024x683px

>>60710638

>>60710919
my sides

File: delid_this_logo.png (334KB, 679x702px) Image search: [Google]

334KB, 679x702px

>>60710638

File: 1493967449274.jpg (510KB, 1315x794px) Image search: [Google]

510KB, 1315x794px

>>60710638

>>60710919
kek

>>60711025
>>60711054
What this thread has to do with Intel? Are amdrones this insecure?

Last two digits will decide its temp

>>60711105
ice cold killa

File: 1494357134227.png (234KB, 882x758px) Image search: [Google]

234KB, 882x758px

>>60711087

>>60711122
Checked for truth.

>>60711105
>>60711122
>>60711132
Proprietary sensor measurement, I guess 5 means 50ºC

>>60710919
This post actually made me remember the absurdity of doing paste for the 6xxx opterons. Thanks. Might need something more than just a plastic card now.

>>60710638
So we are back in the AMD hype cycle that comes after a disappointing release where AMD fanboys start saying, "NEXT RELEASE WILL FINALLY LIVE UP TO THE HYPE THIS TIME! INTEL IS FINISHED!"

File: 1495502393849.jpg (122KB, 750x1024px) Image search: [Google]

122KB, 750x1024px

>>60710977

>>60710977
checked

File: single socket shoah.jpg (78KB, 640x429px) Image search: [Google]

78KB, 640x429px

>>60711087

>>60711195
Ryzen btfo intel though, just look at the latest benchmarks where a 1700x is outperforming a 7700k and is cheaper to boot.

There is literally no reason to buy anything on the intel side of things unless you're destitute poor and can only afford a g4560 - and we haven't seen the R3's yet so intel might yet get btfo there too.

>>60711195
>ryzen
>disappointing
intel shills are truly desperate this time

stop making comically large CPUs and SHIP VIDEO CARDS

>>60711298
Fuck Video cards.

File: 4dchess.jpg (264KB, 1919x1080px) Image search: [Google]

264KB, 1919x1080px

T R A N S C E N D

File: 1469079303705.jpg (78KB, 600x800px) Image search: [Google]

78KB, 600x800px

>>60711195
>6900k perf at a third of the price
>Disappointing

Wut.

>>60711298
You mean for the gaymurz that kept buying NVidia even when they were shipping trash like Thermi? Nah. Fuck them. They don't deserve shit.

>>60711320
Trans-dimensional socket. Nice.

>>60711283
That single core performance tho. The only thing Ryzen beat is the i5, they completely murdered it the to the point Intel had to create the i9 just to replace the i5 with the i7. The 7700k is still the best all around CPU, with nice multi core and awesome single core performances.

>>60711347
After a delid, maybe.

>>60711347
25% better single core isn't a dealbreaker for me, but for some people I guess it is. I'd rather have more overall resources at my disposal.

>>60711347
>delidded corelet stutterfire on a dead socket
>best anything
7700k value is atrociously bad.

>>60711371
It's cheaper than a 1700X, what the fuck are you talking about?

File: IMG_20170526_011713.jpg (4MB, 2592x4608px) Image search: [Google]

4MB, 2592x4608px

>>60711335
"Old de paso advertisement theme plays"

File: 1465622714636.png (727KB, 731x731px) Image search: [Google]

727KB, 731x731px

>>60710660
>>60710743
>>60710770
>>60710919
>>60711320

>>60711435
You're gonna spend the money on a more expensive motherboard for that intel cpu anyway

>>60711490
Well, and you gotta expend more on RAM if you get a AMD cpoo

>>60711087
>Are amdrones this insecure?
Yes

>>60711521
Don't forget the delid kit for your pootel abortion

File: 1245828343472.jpg (4KB, 251x188px) Image search: [Google]

4KB, 251x188px

>>60710919

>>60711371
Only good Intel CPU is the poorfag Pentium.

>>60711347
It's got significantly better perf/watt though which is going to lead to it being able to maintain higher clocks in thermally limited situations like ultra high core counts.
I wouldn't be surprised if the 16/32 threadripper is on par or better than the 18/36 skylake-x in raw performance.

File: Sides in stars.jpg (23KB, 400x288px) Image search: [Google]

23KB, 400x288px

>>60710919
Sides obliterated.

>>60711320
>mfw Zen++ is stacking 7nm dies vertically

File: AMD-Vega-Cube.jpg (157KB, 1366x768px) Image search: [Google]

157KB, 1366x768px

>>60711320
now we know

>>60711821
It will change on how dependent the workload is on single threaded speed. While anything massively parallel should scale better on TR, the possible 4.5 boost from SKLX could be important in completion times of certain tasks. In something like cinebench, TR should crush it mercilessly.

>>60711974
What kind of retard dumps thousands of dollars into a system to run single threaded tasks on it?
You realise this is HEDT right?

>>60712010
single thread perf is literally the only castle intel has
and that is already beaten with the new updates...

>>60711974
Or you could just buy a K series i3/i5/i7 and get even more single threaded performance for much cheaper. Even if it can boost to 4.5, that's probably only with just those one or two tasks going on. If you're heavily using the CPU turboboost won't be nearly as effective.

>>60710648
underrated

>>60711821
I wouldn't be surprised if the 16 core TR outright beats 18 core sk-x. The all-core clockspeed difference is going to be pretty massive.

>>60712010
There are plenty of tasks for HEDT that are dependent on single threaded execution. Not out of poor software, but simply some simulations are heavily sequential. Solidworks particularly comes to mind, where there are a variety of simulations possible so needing a balance of both single and multithread performance isn't unreasonable.

https://www.pugetsystems.com/labs/articles/SOLIDWORKS-2017-AMD-Ryzen-7-1700X-1800X-Performance-908/

>>60710949
Underrated

>>60712117
If 16-core TR has XFR of 4.1 GHz, that's plenty enough. That's barely that far off from 4.5 GHz.

>>60712117
Nigger, the 4 core skylake is outperforming the 10 core broadwell-X in every one of those tests.
How is this a reason for someone to get a 18 core CPU?

>>60712138
Aware, just trying to mention that even real HEDT tasks can be heavily dependent on single thread. I would anticipate most workloads will be better on TR, but there will still be use cases for each.

File: pic_disp.jpg (88KB, 701x549px) Image search: [Google]

88KB, 701x549px

>>60712158
Incorrect.

The simulation and rendering scores for BWLX are higher than the 7700k overall. It's because certain parts of the work are entirely sequential, so single thread speed disproportionately affects completion time.

>>60711347
No, it killed the i7
The i5 was (and is) at a price where it isn't competing with Ryzen

>>60712373
You are forgetting the r5s, which took a giant dump on the entire i5 lineup

>>60711298
Stick to Nvidia you gaymer subhuman

>>60710919
Nicely played anon.

>>60710919
Just apply 1 tennis ball of paste in the middle and let the heat sink spread it

>>60710919
kek

File: delidthis.png (107KB, 2250x2324px) Image search: [Google]

107KB, 2250x2324px

>>60711025
>tfw someone saved your OC
wish you could scale it better, it's 3x bigger in filesize than the original.

>>60713934
10/10

>>60713934
When will they learn?

>>60710977
impressive, very nice

I hope the MacBook Pro refresh goes all out on AMD, that thin Ryzen CPU (Apple said before they used the RX455/460 because of how thin the die was) looked like some kind of hint towards it, not a direct hint but, one could tell.

>>60711087
WHAT THIS POST HAS TO DO WITH
WHAT THIS POST HAS TO DO
WHAT THIS POST HAS TO
WHAT THIS POST HAS
WHAT THIS POST
WHAT THIS
WHAT

File: 1476168760423.jpg (68KB, 754x960px) Image search: [Google]

68KB, 754x960px

>>60710919
Top meme my friend

gonna go full ryzen this christmas, treating myself. pretty hype

>>60714096
If you made that, why did you spell it "this" instead of the more appropriate "dis" ?

>>60715219
so it's a tad more serious. I still have the xcf with sources on my desktop, if I don't forget I'll make the "delid dis" version and post it tomorrow.

File: Kek (2).jpg (39KB, 585x398px) Image search: [Google]

39KB, 585x398px

>>60710919

Well now we are all going to need new coolers

File: massive.png (938KB, 1196x605px) Image search: [Google]

938KB, 1196x605px

>>60715256
noctua's already on it

>>60715120
they will never switch to AMD. They have some agreement with intel on that. Also they have a tendency to put dualcore i7s in the newer MBPs, so go figure.

>>60715280
I think Ill stick with Intel

File: muh cores.png (353KB, 915x678px) Image search: [Google]

353KB, 915x678px

>>60715288
>MacBook Pro
>Pro

File: AaB5Gp7.jpg (155KB, 662x466px) Image search: [Google]

155KB, 662x466px

>>60715280
Apply directly to the forehead.

>>60715280
that's a really really big poo fan cooler.

>>60715324
Bigger surface area = better heat dissipation. There's also seperate dies under the lid. It's definitely not a bad thing that it's this big.

>>60715331
oh yea, it's just called Macbook now? I stopped following this shit when they made it stupid retarded with naming.

>>60715356
Ya but to upgrade I will need a new motherboard case and ram at least and the case will need to be massive

>>60715368
Not really. Just because the baseplate is bigger doesn't mean the cooler and fan are enormous and won't fit, unless you've got some tiny mITX case or something.

>>60715393
The motherboards will have to be larger or cut features

>>60715368
those are 120mm fans, so it's smaller than a dh-15

and that cooler was designed to fit in a 4U server case, it WILL fit in literally any non trash case on the market

File: ASRock-X399-Taichi-and-ASRock-X399-Professional-Gaming-740x555[1].jpg (69KB, 740x555px) Image search: [Google]

69KB, 740x555px

>>60715404
yeah... no, you're stupid

File: Gigabyte-X399-AORUS-Gaming-7-AMD-Ryzen-Threadripper-Motherboard_1.jpg (195KB, 800x534px) Image search: [Google]

195KB, 800x534px

>>60715404
No, they are standard sized. It is a bit of a tight fit, though.

>>60715427
Other than being from asrock that doesn't look too bad

File: Gigabyte-X399-AORUS-Gaming-7-AMD-Ryzen-Threadripper-Motherboard_2[1].jpg (86KB, 800x534px) Image search: [Google]

86KB, 800x534px

>>60715442
Gigabyte one pic related and >>60715437

>>60715448
I hate that I might have to go AND soon we will see what i9s are

>>60715458
translate pls

>>60715288

>Also they have a tendency to put dualcore i7s in the newer MBPs, so go figure.

In the 13", yeah. 15" always gets 45w quad cores.

File: bigcpu4u.png (270KB, 770x578px) Image search: [Google]

270KB, 770x578px

>>60710648

>>60715479
Replace and with AMD

>>60710977
did anyone else look at the thumbnail and think 'Hey Arnold'?

>>60715458
>>60715496
well, AMD built a nice architecture this time, there is no shame in using it instead of intels shittier one

>>60715516
Like I said I'll wait and see what the i9 turns out to be

File: i9.png (2MB, 1920x1200px) Image search: [Google]

2MB, 1920x1200px

>>60715527

File: naplespackage.png (1MB, 1024x775px) Image search: [Google]

1MB, 1024x775px

Dont forget Epyc will have 4 dies on the same package.

>>60715510
kek

File: 07062016-AuburnFire01-780x585.jpg (73KB, 780x585px) Image search: [Google]

73KB, 780x585px

>>60715527
https://www.overclock3d.net/news/cpu_mainboard/intel_s_skylake-x_and_kaby_lake-x_cpus_will_not_be_soldered/1
http://www.gamersnexus.net/news-pc/2936-intel-i9-7900x-delidding-cpu-package-thermal-paste

>>60715553
You're a big die.

>>60715486
retarded if you ask me
you pay extra to get a premium experience and you get a fucking dualcore throttling to shit

File: Rosewill_Server_Chassis_Server_Case_Rackmount_Case_4U_Metal__38_res.jpg (185KB, 834x500px) Image search: [Google]

185KB, 834x500px

>>60715562

>>60715527
intel shot themselves in the foot with that thing, they will only get good performance with those designs if they give a big fuck you and go 200W+ tdp

>>60715368
>I will need a new motherboard case and ram a
You will also need a new motherboard and DDR4 ram
Most importantly, AMD doesn't release new sockets every fucking year

>>60715368
you can wait(tm) for other cpu cooler brand

>>60715593
I'm going to wait to see what the best option is the last memory I have of AND is failing miserably at ps2 emulation

File: 1392934192095.gif (2MB, 384x216px) Image search: [Google]

2MB, 384x216px

>>60715573
when you buy Apple, you are guaranteed to pay a premium for a throttling device.

File: IMG_1088.png (123KB, 240x286px) Image search: [Google]

123KB, 240x286px

>>60710919
EBYN :DDDDD

>people think engineering samples aren't soldered so the real thing won't be too

Just wait, AMD shills.

File: clocksoc.png (1013KB, 960x686px) Image search: [Google]

1013KB, 960x686px

>>60717009
>Intel is now Just Wait company
Oh boy.

>>60717009
just like kaby lake was soldered?

>>60717009
>engineering samples
People in the know like de8auer wouldn't take the time to make a delidding tool for it if it was just the engineering sample

Or are you basically admitting Intel are woodscrews tier now.

File: 1496427478955.jpg (87KB, 500x872px) Image search: [Google]

87KB, 500x872px

>>60717009

>>60711087
No but sounds like (you) are

File: smug_lisa_su_face.png (355KB, 777x656px) Image search: [Google]

355KB, 777x656px

>>60717009
>Just Wait(TM)

>>60714096
good shit. Looking forward to delidding after warranty runs out.

No vm support yet. Ill stay on the sidelines.

Great material for a dual socket homelab.

>>60717116
They fixed some of the VM shit. IOMMU groups work now.

>>60717116
AGESA 1.0.0.6 fixes some VM-related bugs and adds PCI-E ACS support.

>>60715553
Tech retard here, each die is 8 cores meaning Summit Ridge/1800x? That's pretty cool. Would stacking the dies vertically help in any way aside from package size reduction?

>>60717211
Yes and no.

>>60717116
Wasn't Ryzen updated already? I'm surprised Wendell hasn't had an orgasm yet.

>>60717211
if you stack them vertically the traces would get fucked, because they would either have to circle around the dies in the center or below them

>>60717151
>>60717197
>>60717233
Theres plenty of hacks floating around and of course there'll be proper support sooner or later. But by that stage i reckon we'll on another generation.

Ill wait and see.

>>60715337
Underrated

>>60710770
Big Beautiful CPUS

File: 1482783597496.png (346KB, 500x464px) Image search: [Google]

346KB, 500x464px

>>60710638
>they're actually going to sell an enterprise(tm) targeted chip under the name "threadripper"
please gas AMD's marketing department that sounds terrible and triggering

>>60717748
> "core i"
> anything between a dual core celeron and a 16-core pentium

>>60717748
No, the server CPU is Epyc. Get your embarrasingly stupid names right.

So when are we getting cpu die floor tiles?

>>60717748
>core i9
We're talking cringe huh.

Also Xeons bling-bling edition is even more cringy than EPYC. At least EPYC serves Intel a grim reminder.

>>60717827
>>60717850
>>60717859
>how dare you not lick the boots of my favorite rich kike corporation, you must obviously be white knighting for the other one!
consider suicide
>>60717836
that's way better then, I thought this was the server part

>>60717909
Are you fucking retarded? XEON PLATINUM (HOOKERS INCLUDED) beats everything on cringe.

>>60717922
>MFHFHFMFMF BUT INTEL (sucking continues)
are you a terrible troll, a paid shill or just really this much of a boot licking brainlet that you literally cannot comprehend the idea of someone existing that doesn't suck either AMD or Intel cock at every opportunity? they're both shit names dude

>>60715256
there was a vid about that saying youll be able to use normal coolers since you only need to keep the dies cool and they are close together.

>>60713934
kek

Intel is going to be behind for the next 8 years, again.

Goddammit i dont want the amd cpu + nvidia gpu meme to come back

>>60710949
Nice

>>60715573

Well, nobody would ask you. because you think its smart to put a 45 watt CPU in a chassis with 1 fan instead of 2. Unlike you, I've used both (touchbar models) extensively. I'm yet to throttle.

>>60715331

Lol. I like how it now includes 16 cores.

File: maxresdefault.jpg (180KB, 1920x1080px) Image search: [Google]

180KB, 1920x1080px

>>60715492

File: my imagination is too poor to think of a name for this masterpiece.png (49KB, 143x157px) Image search: [Google]

49KB, 143x157px

>>60710977
>>60715077

>>60711320
Is this like a shitty /k/ube?

>>60715331
>>60718760
16 cores is arguably getting into diminishing returns already.
On the other hand
>18 gores
>having buy another Windows license for those 2 extra cores
lel

>>60717211
also how to remove be heat.
can't get it away from dies easily if stacked

>>60719541
>using Windows
Kek

>>60719666
>actually
>really
>unironically
>installing Gentoo

>>60718905
>starte
you tried

File: 1496421497627.jpg (151KB, 725x512px) Image search: [Google]

151KB, 725x512px

>>60710638

File: jewshown.png (138KB, 354x504px) Image search: [Google]

138KB, 354x504px

>>60710638
>AMD JewRipper

File: 1496328708889.png (745KB, 1260x673px) Image search: [Google]

745KB, 1260x673px

>>60717009
Even LN2 can't help this.

>>60720127
>10c/20t
cute

File: 1441428115295.jpg (181KB, 563x821px) Image search: [Google]

181KB, 563x821px

>goy, you don't need to solder your cpu to your IHS

File: 1348635013249.jpg (21KB, 283x300px) Image search: [Google]

21KB, 283x300px

>>60717500

>>60717062
Soldering small modern dies is not recommended

>>60720629
Yeah, let's use some tooth paste instead.

>>60720641
You can use good quality tim instead of solder and get decent results still, just Intel didn't do that and also fucked up their ihs glue application

>>60711435
The 1700 is cheaper, and it's pretty much the same chip.

>>60721205
Binning

>>60720327
Holy kek.

>>60720127
>LN2
>only 4.4GHz
This is a disaster.

>>60720327
There is a meme here that I'm not understanding.

>>60721491
You misread. I considered allowing such a misconception continue. But that's not fair, because it's still incredibly shit even tho it wan on an AIO.

>>60721525
>Core voltage: 1.764 V
No fucking way.

>>60721542
To force it to throttle and thus fake the Ghz

>>60720127
>Single thread 527
>Multi thread 5198
How ? My 4690k pulls around 2100/8000 at 4.7GHz.

>>60721604
Update nigga.
They changed the scoring system.

>>60721612
Oh okay.

File: 1483785721324.png (429KB, 451x619px) Image search: [Google]

429KB, 451x619px

>>60715337
>>60717500
>>60720400
>not using the builtin water cooling of your body

>>60715359
No, it's still called MacBook Pro. But now the current generation MBP isn't fit for any "pro" tasks beyond simple text editing.

>>60720076
best one

File: delidDIS.png (109KB, 2250x2324px) Image search: [Google]

109KB, 2250x2324px

>>60715219
here's the "appropriate" version

>AM4 will last a decade, I swear
>release new socket in a matter of months

>>60715527

It's gonna be expensive as fuck is what it is going to be.

>>60713934
manlets when will they learn ?

>>60710919
10/10

>>60721989
>being this rertarded

>>60711320
this is the real 3d xpoint technology

>>60710648
fpbp
kek

File: 1493283204085.jpg (61KB, 500x667px) Image search: [Google]

61KB, 500x667px

Do I smell a penny?

>>60722552
What the fuck

>>60711025
>>60711054
>>60711130
lotta loyalty for a corporate consumer

File: 1490640721096.gif (1MB, 360x354px) Image search: [Google]

1MB, 360x354px

>>60717748
>oh noes, why they call it like that, I no like it
Why the fuck you care? You're using name or what?
Oh wait, unless you're brandboi, then of course you'd care about every single detail!

p.s. you must be 18 to post here.

File: 1486037815818.png (216KB, 568x1023px) Image search: [Google]

216KB, 568x1023px

>>60721989
No one said it will last a decade. Mostly I read people are saying it will last at least 4 years.
New socket is for server CPUs.

I hope you feel embarrassed.

>>60722723
I meant
>non consumer CPUs, but for high end desktops (production machines)
not server... tsss...

>>60722785
Both theadripper and epyc sockets are the same thing.

>>60711453
oh, the stock cooler has neat LEDs? I bought a water cooler for nothing then...

>>60711347
>create
>i9
>implying it isn't just a renamed i7x

>>60711322
amada a cute

>>60713934
holy fug

>>60711087
>so buttmad he can't sentence right

File: 6i3IDK2.png (249KB, 623x767px) Image search: [Google]

249KB, 623x767px

What a good time to buy CPUs

>>60710638

CPUlets can't into thread raping.

Can someone explain why this unit is so damn big? Is it using old tech?

>>60725210
iirc it has 2 physical packages inside.. I might be wrong and some autist will correct me


also it has over 4000 pins and 60 pci-e lanes which all need space

>>60725210
It's literally two CPUs glued together, so it's going to be at least twice as big. Also, it's based on the server socket which can support quad-die chips, so it's actually over-built for the intended purpose.
The neat thing is that we may even be able to slot in Epyc chips the same way Intel lets you put Xeons into their enthusiast boards.

File: 1445146298432.jpg (137KB, 620x720px) Image search: [Google]

137KB, 620x720px

>>60717850

>>60725402
Wonder how she'd look covered in jizz TIM.

File: ryzen three d ripper.jpg (111KB, 1024x680px) Image search: [Google]

111KB, 1024x680px

>>60711320
Got me inspired

>>60725439
How many stacked dies you got in there? A billion?

>>60713934
>right == me
>left == that guy she tells you not worry about

File: 13669003_292353274441732_8443115734833796016_n.jpg (21KB, 691x597px) Image search: [Google]

21KB, 691x597px

>>60725439

>>60725210
They decided to use the same socket as their server CPUs for threadripper, which needs all the pins for 128 PCI-E lanes and 8 memory channels. On threadripper most of the pins are redundant.

>>60725466
it's better to use the same socket rather than create more separate ones. plus better cooler compatibility for servers and desktops too I guess...

>>60715406
Serious question none of the current aio coolers will do, but considering that you actually don't need to fit the whole package, since the heat is spredading it could work with just a different mount.
Another thing to consider, with such a big service area, this CPU can be really cooled effectively, even using a small cooler, since the thermal transfer will be optimal and soldered.

>>60715581
Tdp is one thing another is the incapacity of properly transferring heat,which eliminate one of Intel advantages, overclocking, and even prompt thermal throttle on regular usage.

>>60717009
No idiot these are already production runs.

>>60721516
happy merchant nose man

File: TR.jpg (135KB, 1500x659px) Image search: [Google]

135KB, 1500x659px

>>60725439
>>60725702
Made yet another new thing

File: 1493909669592.png (495KB, 600x572px) Image search: [Google]

495KB, 600x572px

>>60725410
delet

File: MayoLake.jpg (165KB, 1000x1000px) Image search: [Google]

165KB, 1000x1000px

>>60725814
>>60725820
I'm also the anon who made this

>>60725820
>9 watts
>9 Ghz
What kind of hax put this thing together?

>>60710919
kek

>EBYC/EBIN or whatever is the server CPU
>Threadripper is supposed to be standard consumer CPU
>800 dollary doos

What the fuck are you going to do with an 800 burger CPU? No vidya is going to need it and you can do a workstation CPU for half of that.

>>60725903
>and you can do a workstation CPU for half of that
Used xeons aren't going to be worth it for long after this.

File: 1849892.png (1MB, 1920x1080px) Image search: [Google]

1MB, 1920x1080px

>>60720127
k

>>60725992
>1.557v
>Ln2
WOW

>>60725820
>not L9 cache
>not chipset X99
Cirno makes my peepee feel weird tbqh

File: IMG_5785.png (7KB, 531x232px) Image search: [Google]

7KB, 531x232px

>>60725992
>Has to be OC'ed on LN2 just to break 3100 points
>TR can hit over 3000 stock
Into the trash it goes

>>60726120
Thats the 10 core intel chip, not the 18

>>60726120
That score is fake and honestly I'd be disappointed by it anyway. I'm expecting 16 core TH to hit somewhere between 3350-3450

>>60726140
Which is still going to cost like what $1000? If the rumored TR price is true, it's going to be quite a bit cheaper than even the 10-core.

>>60725439
Glorious AMD steel folded over a million times

>>60726120
But anon, it SHATTERED the previous world record... !

>>60725903
Because CPUs are only used for video gay memes

>>60725903
if you don't know what it's for, it's not for you.
go away.

>>60726354
Are kikes censoring us?

>>60726430
No he realized he was attacking his own side.
Score shown: 3000
Anons hope: 3350

He is claiming fake due to the score being lower than thought.
Also the score confirms that LN2 was used.
http://hwbot.org/submission/3563150_elmor_cinebench___r15_core_i9_7900x_3181_cb

>>60726487
>>60726120
>>60726149
I'm thinking TR will be faster than that even. Didn't the blender benchmark show more than 2.5x scaling?

>>60726833
I'm remaining cautiously optimistic on IF scaling across dies.

>>60726833
>>60726858
even if it gets just around 3000 for the 16 core, it still costs the same as intels 10 core which needs ln2 to achieve this score, and it has 20 more pci lanes too

>amd
>that size
40ºC on idle confirmed.

>>60726980
that's not an intel processor tho

>>60727019
>superior product
>double the price
>toothpaste
>dogshit clockspeeds
>laughable amount of pci lanes

>>60710638
I don't know why there are so many threads about this, how many of you are actualy going to get a 16 core cpu? A tiny fraction i'd imagine

>>60727010
Of course it's not

>>60726980
The massive size of the thing means better heat dissipation, not worse. And the dies are actually soldered. And it has lower TDP.

>>60726833
>>60726858
Something tells me quad-channel memory has a hand in this, perhaps the higher memory bandwidth offsets the limitations seen in the 8-core part?

>>60726980
Larger surface area is actually an advantage, because it's easier to dissipate heat, especially with multiple discreet dies rather than a single monolithic die ala Intel.
If the architecture allows it, a larger die is usually better.

>>60727049
>rebranded xeons
>singlethreaded performance
kek

>>60710743
kek

File: muhsinglecore.png (453KB, 1920x1080px) Image search: [Google]

453KB, 1920x1080px

>>60727049
>giving a shit about MUH SINGLE CORE when talking about HEDT
>not realizing just how tiny that single core advantage is
kek.

File: __takagi_san_karakai_jouzu_no_takagi_san_drawn_by_yamamoto_souichirou__sample-5ad1b8b859c6a4f73eeea37aea35aac3.jpg (103KB, 850x1200px) Image search: [Google]

103KB, 850x1200px

>>60727019
>Genuine Intel Toothpaste
>superior product

>>60710919
nice

>>60727019
Dumb frogposter.

File: Ryzen.jpg (46KB, 1024x680px) Image search: [Google]

46KB, 1024x680px

>>60725439
built in water cooling?
that's what I call ingenuity

>>60727105
lol. It doesn't really need watercooling, though.

>>60727116
>cpus are only useful for games
>>>/v/

>>60727105
>built in water cooling?
That was already done before. But within the CPU itself.
I forget the brand. Should be related to the quote.
'apple used the CPU to design the next Mac, so I used the Mac to design the next cpu'
Something to that effect.

>>60727165
>heavy duty
>singlethreaded workloads

>>60727128
Ironically, Threadripper will have better single-threaded performance than Ryzen 5 or 7 anyway due to the 2:1 infinity fabric speed.

Why is it so hard for intelfags to accept that they lost?
Why do they care so much?

>>60727148

Lol. Shills this desperate.

Enjoy paying per core by the hour!

>>60727272
didn't know professional shills had that

>>60727272

>His shilling is this desperate that he thinks a workstation level workload requires HPC time.

>>60727019
>>60727049
>>60727116
>>60727148
>>60727165
>>60727272
>>60727329
Reminder to report shitposters.

File: super nut.jpg (15KB, 480x475px) Image search: [Google]

15KB, 480x475px

>>60725439
>1 gorillion cores

File: girls-laughing.jpg (55KB, 553x496px) Image search: [Google]

55KB, 553x496px

>>60727387
>MFW all those shitposts were one person.

File: __amada_kokoro_amd__08123da26e693fc716b23224ca67cff8.jpg (192KB, 1706x2048px) Image search: [Google]

192KB, 1706x2048px

Holy fuck, those rumored clockspeeds are really making me consider going HEDT this time around. Tempting, but I think I'll go "modest" and grab a 1700 this time around since it's even cheaper now.
What are you doing this time around? Thinking it through, I think it would be best to go modest with the 1700 and then go all out with 7nm ZEN+ next year.

>all that (Dead)

thank you Jani

>>60710638
its not big boy. its ebola - T%%readT%%RIP%%P%%ER

>>60727696
t % ri % pp & 4 &

thanks based janitors/mods

File: smug-pupper.jpg (365KB, 1057x1008px) Image search: [Google]

365KB, 1057x1008px

>>60710977
>kikeripper

>>60725402
>>60725820
Cirnofags should end themselves tbqh

File: 1495910945016.jpg (112KB, 750x746px) Image search: [Google]

112KB, 750x746px

>>60727105

Make better, please.

>>60727828
can't be arsed
my mouse double clicks or doesn't click at all which makes day to day tasks a pain, and I don't really have much image editing skills

File: 1494894361219.jpg (75KB, 425x516px) Image search: [Google]

75KB, 425x516px

>>60727900

I send a new mouse for you

File: XpOtynp.jpg (79KB, 800x600px) Image search: [Google]

79KB, 800x600px

>>60711320
Boy, I haven't used this image in a while

File: 1496501710995.jpg (90KB, 1024x680px) Image search: [Google]

90KB, 1024x680px

>>60725439
Rate my fingerbox

>>60728197
10,000 lightyears in MS paint?

>>60725903
>>Threadripper is supposed to be standard consumer CPU
Threadripper is as much of a standard consumer CPU as Intel's HEDT lineup.

Kikeripper is real lmao

>>60727063
Anon no a bigger die is hotter

>>60725402
>>60725820
>implying Cirno wouldn't go with AMD super sekrit 1999XX for the highest binned thread rippers
Embarrassing desu.

>>60712201
That image is 3 months old. I'd like to see Ryzen comparisons from the last few weeks.

File: 1493895329997.gif (375KB, 480x480px) Image search: [Google]

375KB, 480x480px

>>60725439
>>60727105
>>60728197
Think of that like samsung's 3D NAND - building a multi die CPU with space between the separate dies for interconnects AND channels where the coolant will flow through. so instead of placing 4 dies next to each other you could stack them for the cooling efficiency. I know that's practically impossible from semicon manufacturing process because of the timing issues (link lenght etc) but wouldn't that be interesting?

>>60728522
This.

File: epyc.jpg (44KB, 957x539px) Image search: [Google]

44KB, 957x539px

Are you ready for the 88-core model?

>>60717211
It would help reduce the latency and power consumption of their interconnects

>>60728984
and would make cooling impossible

>>60728221
yea

>>60728982
That's a very 'shoop.

>>60728982
nice

File: 1283018399825.jpg (218KB, 600x489px) Image search: [Google]

218KB, 600x489px

>>60728982
here's a mobo to go with it

>>60710638
What is that thing called where they stitch the CPU together or whatever
quantum stitching?
magic thread?
I fucking forgot

>>60729172
Multi-chip module, or MCM.

>>60725451
nice facebook meme faggot

File: assymmetry.jpg (87KB, 725x512px) Image search: [Google]

87KB, 725x512px

>>60710638
Only now I realise AMD tends to one of my biggest gripes with LGA2011. While Intel chips can be mounted the wrong way without you noticing before the magic smoke leaves them, AMD bothered to install safeguards against this.

>>60728982
But anon, that's only 56.

>>60729413
Dude LGA 2011 has those too

>>60729495
But they're symmetrical, you can turn the LGA2011 around if you want. Close the socket without noticing anything strange.

>buy server parts from server shoppe
>get instructions to triple-check my installation twice before powering on
>guy notes that he managed to fry an expensive mobo and CPU destined for a customer

File: big-don-lil-yeb.jpg (488KB, 1024x1536px) Image search: [Google]

488KB, 1024x1536px

he is a big boy :D

keep eating chub chub!

you gonna be big n stronk one day

File: Screenshot_3.png (1MB, 1843x1039px) Image search: [Google]

1MB, 1843x1039px

>>60729578
Not true at all
They are actually offset from the top, symmetrical on the sides but they aren't in the middle of the chip. It's not possible to mount them incorrectly
Red is longer than green

>>60729644
yeah but you can still mount it wrong if you aren't paying attention
sure the processor won't go in all the way so if you clamp it down it will break the PCB

File: 1495819893801.png (306KB, 345x366px) Image search: [Google]

306KB, 345x366px

>>60728982

>>60729677
You just described every processor with these notches

>>60710977
Can't read in reddit, sorry.

>>60729861
>janitors delete dissenting opinions about AMD
>keep a gets post outside of /b/
they are fucking retarded

>>60729172

Infinity Fabric.

>>60729980
>dissenting opinions
Yeah, that's not what that shitposter was doing. At all.

>>60729439
it's 112, each die is an 8core

>>60730357
>dissenting opinions are shitposting

>>60730443
>deliberate bait and lies are dissenting opinions

>>60728982
If IF could actually scale like that I don't think there'd be anything left of Intel

>>60730589
going by rumors, zen 2 (7nm) will have 6 core CCX, thus allowing the highest end server cpu to have 48c/96t on a 4 die configuration, it's not 88 cores or a fuckton of dies, but it's a start

>>60730630
Going by density at 7nm AMD would have no problem putting 8 cores per CCX

>>60730667
>Going by density
Source? Just because 7 is half of 14 doesn't mean it has double the density, especially when these nm figures are completely meaningless now.

>>60730667
they will probably keep at "just" 6 and also improve upon other things, like clockspeeds and efficiency

a 50+% core count bump on all cpus is enough from one generation to the other, I think

>>60730689
>>60730689
Based on FinFET transistor architecture and optical lithography, with EUV compatibility at key levels, the new 7nm FinFET technology is expected to deliver more than twice the logic and SRAM density and >30% performance boost compared to 14nm foundry FinFET offerings. This technology will be supported by a full platform of foundation and complex intellectual property (IP), including an application-specific integrated circuit (ASIC) offering.

https://www.globalfoundries.com/technology-solutions/cmos/performance/7nm-finfet

Density is counted squared, 7nm^2 is a lot smaller than 14nm^2


In theory.

>>60730550
>it's ok when AMD does it

>>60730789
feel free to show examples

>>60730805
>literally just making shit up saying vega runs 180 fps in prey and skylake X is going to be a housefire because they use thermal paste

>>60730696
these are the figures on zen 2 if this is true:

Zen 1 (14nm) --> Zen 2 (7nm):

>Ryzen:
4c/8t --> 6c/12t
6c/12t --> 10c/20t
8c/16t --> 12c/24t

>Threadripper:
10c/20t --> 16c/32t
12c/24t --> 18c/32t
16c/32t --> 24c/48t

>Epyc:
16c/32t --> 24c/48t
32c/64t --> 48c/96t

>>60730819
AMD did this when?

>>60730819
>vega runs 180 fps in prey
That guy is an idiot, but at no point did he start trashing NVidia in the process
>skylake X is going to be a housefire because they use thermal paste
7700K is already hot if you try and OC it w/o delid, and SK-X uses even higher TDP, so I don't see how this is lying

>>60730865
the 4790k used thermal paste and wasn't hot
6700k used thermal paste and wasn't hot
Literally just the 7700k
>w-well he's an idiot
excuses

>>60730819
>polite speculation
>previsions based on current evidence of thermal paste performance

none of these are similar to
>/v/ fag claiming that the HEDT platform should perform well in poorly threaded games, and if you need to do high threaded workloads you should rent a server farm
>some idiot claiming the intel 18 core could run all cores at 4.8 GHz on water
>some other idiot trashing amd for literally no reason other than butthurt

>>60730667
Too overkill, remember that their APUs are made out of 1 CCX, and there's no way dual channel can feed 16 cores if they want to keep AM4 compatibility.
Maybe AM4+ for 18 cores, but AM4 likely can't go over 12 cores unless mobos are routed for 1DPC for 4dimm/4 channel memory.

>>60730885
>things AMDumbfucks do all the time but are ignored
You realize everyone is a fucking idiot right? Intel fanboys, AMD fanboys, all fucking retarded. However you guys try to defend each side as if it matters. It's pathetic really

threadripper sounds like what occurs in pajeets pants after a vindaloo

>>60730823
Threadripper is 8, 12 or 16 cores. Nothing else.

Epyc is 16 , 24 and 32.
Zen SMU doesn't allow non-symetrical configurations.

>>60731033
I really doubt that there is an 8 core threadripper, if there isn't a 10 core, then it will be 12 and 16 only

>>60730823
You took all the time to list the c/t, when you could have just said "with SMT" but yet you imply that the 18 core varient has SMT turned off on 3 cores. And one of those cores doesn't work at all.

>>60731094
You think? Twice the channels and cache on cheap 8 core at crazy stock clocks sounds very interesting.

>>60731210
sorry, I did this yesterday late night, didn't pay enough attention
>>60731212
there won't be crazy stock clocks, only lower voltages and clocks similar to the 1800x

>>60731238
For an 8 core at almost twice the TDP? base/turbo clocks over a 1800X would be higher, with the benefit of quad channel IMC and 36MB of L3

>>60731265
32MB

>>60731265
but the voltage/frequency wall is still there since those dies are basically the same, ryzen refuses to post if it's above 4.1 GHz

>>60731265
We're looking at B2, not B1 silicon, that might change.
And 1800X is still 500-600MHz from its limit at stock.

>>60731376
Meant for
>>60731338

>>60731398
maybe we could see some small improvements in max clocks with B2, coupled with higher stock clocks, but don't expect more than 100~200MHz

>>60731455
>>60731455
It could also improve the voltage curve, ryzen silicon was pretty rushed as is.

>>60731513
ryzen silicon is currently built with a process that doesn't allow for great clockspeeds unfortunately

this will possibly change with zen 2

>>60730667
>Going by density at 7nm AMD would have no problem putting 8 cores per CCX

Best case is 2x 6c CCXs.
> shared 12MB L3 victim cache
> more inter-core traffic can go through ~100 GB/s local L3 instead of 40-50 GB/s IF link

Worst case is 3x 4c CCXs.
> still only 8MB shared L3 available to any given core
> per-CCX IF link gets even congested
> more snoop traffic clogging up IF links since probes need to be sent to every L3 in most cases (good filtering can reduce this somewhat, but more L3s = more space needed for filter tables)

8c CCXs simply aren't happening, since crossbars (whether internal L3 or external IF) don't scale to high port counts very well, and Zeppelin would already have been a single 8c CCX if the design was very feasible.

>>60731602
Yes, but both the process backend and silicon can be tuned to get something higher.
The assumption that LPP is hardlocked at those clocked is flawed, a lot of work goes in there.
Good example is upcoming Vega actually, LPP at 1600+ Mhz, while polaris shat itself at 1400MHz

>>60729172
Unyielding thread.