THANK YOU BASED NVIDIA

>Maxsun chin chon
>only 1 8pin
Ofc, if you remove half the VRM you should remove half the power supply to not make that chinkshit explode.

What is your point exactly?

Why pooVega did not have custom cards at the launch?

It's chink shit, literal trash. There's nothing to see.
Also what the other user said, no specs.

Almost all cards are made in China m8

based

Attached: Untitled.png (751x799, 687K)

There's a difference between some unknown chink brand and reputable brands.
Go look up cheap chink GPUs with VRM that will shit itself.

>reputable brands
Unironicaly chinese AIBs trends to put more vrm on the PCB and Anus, MemeSi, Niggabyte, Evegay all use chinese vrm/component. The difference is "chink" brands are not available in Europe/America. They are not cheaper.

>he actually believes this
Enjoy your chink shit housefire.
>chink components
Designed in China and made in China are two different things.
Asus for example uses APAQ caps.

And reputable chinese brands also use high quality components. What is your point? Are you on a quest of All Chinese Brands Bad All Non Chinese Brands Good?

>housefire
But you told me that ()
>APAQ caps
Create in Taiwan, made in China

Autism: the post

Attached: 1529504013270.png (1024x595, 559K)

>videocardz

Attached: videojewz.png (967x453, 39K)

videocardz.com/newz/amd-ryzen-threadripper-2990wx-available-for-preorder-for-1799-usd
here's the link

Why the hell do you talk about AMD CPU in a GPU thread?

Who buys $2k cpus which are useless fo anything professional apart from video encoding? Nope, you need xeon/epyc if you do CAD work fo a living.

why the fuck do they jump from 11xx to 20xx

Attached: bombing.jpg (494x485, 31K)

>$1800
>no ECC
>not a niche

They will use Epyc on servers, not Threadrippers

I hope you realize that Maxsun is one of the top three brands in china and that they ship more GPUs in just china than EVGA does globally.

Fucking Maxsun GT1030s account for 20% of all Nvidia Geforce sales volume.

I do not fucking care
I only buy ASUS cards since they are the BEST.

Expensive, lower performance expected, HBM

Nah, they just spend a lot on their ROG gamershit branding.

Palit/Gainward which owns Daytona, Galaxy/Galax, Vivkoo, Yuan, and XpertVision while also doing OEM rebrands for many other companies ships more GPUs than Asus, occupying 20~25% of the world's graphics solutions.

It's not confirmed yet

you must be 18 or older to post here

>using less of the same doesn't make any difference
Ok chin chon, enjoy housefire.
>Create in Taiwan, made in China
That's the point. It's not created in China.
If it was created in china, it would be dogshit like anything else.
Nice argument, ping pong.
>producing more makes them better
Lmao.

Ryzen and TR platforms both support ECC brainlet. A $1800 is niche. I personally would invest in a proper epyc platform for that price. I actually am a TR1950x owner.

I will likely be buying Nvidia's new GPUs to resell on ebay for 50% over MSRP. I don't think they'll deliver value in the upcoming line.
> No Nvlink
> No open access to the new hardware that will support Real time ray tracing
> Still no open access to nvenc

So fuck Nvidya. The current cards are more than sufficient. I'm waiting until 7nm and a point when both AMD/Nvidia finally open up their platforms and enterprise features.

>Ryzen and TR platforms both support ECC brainlet.
>what are buffered and unbuffered ECC

The 2080 is reportedly going to be priced at $1500

Attached: 1448862398635.jpg (263x383, 41K)

...

Do be fooled, you really think because it's from Taiwan aka China it's better? Oh boy

>Taiwan aka China
Lil chin zhu xo missing the old days? Lmao.
Yeahs it's better, just like with japanese vs chinese components.

No, $2999, nVidia BTFO. Team red FTW XDDD

But both use Chinese components Jesus fuck one is not better than the others

>using less of the same doesn't make any difference
And how do you know chink cards use less components I wonder? Surely you have a way to back up that claim, don't you?

>Intel's 28c won't compete too well
Oy gevalt, clam your anti-Semitism, goy

Attached: 1499959791737.png (650x462, 697K)

this, why did i click...

works for AMD, now nvidia is going to find out how well it works and do it all the time too.. :(

Nope. VW made in China is still not chinese, it's german, just as APAQ are taiwanese.
Less phases. Go to YouTube and see for yourself.

> What is the chance you get an actual bit error
Slim to none
> ECC is a fucking meme that only has relevance in a small range of enterprise operations.
> Buffered/unbuffered have to do w/ easing electrical load and proper scaling for systems w/ 1TB of ram YOU FUCKING BRAINLET

Why do poorfag core-lets always dream of stupid ass useless features they don't need? fuck off w/ your ecc meme and definitely fuck off w/ unbuffered/buffered.

Broke ass poorfag probably running a core2duo on chink tier DDR3 :
> MUH BUFFERED ECC RAM
dream louder kid

Post the link please

>Maxsun 1070 housefire 4 phase
youtu.be/xxbf_zq5YMQ
You can clearly see 1/3 of the VRM missing.
>Asus 1070 Dual low end model with 6 phase quality VRM
youtu.be/Xm4aBe23kec

>Maxsun 1070 housefire 4 phase
Exactly like the Nvidia ref, what is your point????

Attached: front_full.jpg (4872x1950, 2.53M)

Attached: Screenshot_20180806-102027.png (1920x1080, 1.18M)

>VW made in China is still not chinese, it's german
VW made in China is engineered by Krauts and put together by Chinks- it's not really anything. In much the same way that a Volvo assembled in Belgium in a plant ultimately owned by a Chinese company is not really any one nationality either.

videocardz.com/76953/nvidia-geforce-gtx-2080-1180-pcb-has-been-leaked

THANK YOU BASED NVIDIA

REMOVING OUTDATED, OBSOLETE LEGACY DVI SHIT FROM PCB & GPU

NO LONGER WILL SHITTY TAIWANESE & CHINA COMPANIES BLOCK THE EXHAUST SLOT WITH DVI CONNECTOR

Attached: NVIDIA-GeForce-GTX-2080-1180-PCB-1.jpg (2000x979, 545K)

Old new...

Garbage PCB and shit for overclocking. That is why only Nvidia fanboys and retards buy FE cards instead of custom ones.
>not really any one nationality either
This is where you are wrong. It doesn't matter where it is manufactured. Volvo is and always will be swedish same as VW is and always will be german.

>Garbage PCB and shit for overclocking. That is why only Nvidia fanboys and retards buy FE cards instead of custom ones.

You clearly don't have a pascal GPU. Reference card are GOAT. KYS

Attached: Screenshot_20180806-103028.jpg (659x296, 78K)

That's the low tier 1070 basically reference PCB with an open cooler for people who want a cheap 1070 to slap a water cooler to.

Custom PCBs are bitch for water cooling.

Oh fuck, they removed the DVI port?
How the fuck will I connect my 5 devices now if they took away a port.

I don't want to buy a second card.

>Garbage PCB and shit for overclocking. That is why only Nvidia fanboys and retards buy FE cards instead of custom ones.
The overckocking on reference card are better with Pascal GPU retard, stop live in 2015. People hate it because they make more noise.

>Why do poorfag core-lets always dream of stupid ass useless features they don't need? fuck off w/ your ecc meme and definitely fuck off w/ unbuffered/buffered.

Unironically kill yourself.

You want moar proofs :^)

Attached: Screenshot_20180806-103753.jpg (968x604, 205K)

>clearly worse VRM
>i-it's just as good goy, buy FE!
Sure thing you cuckboys
>Screenshot
Nice source.

Which company uses the best components for maximum overclockability and performance?

>The mini gets the best clocks and is also the cheapest of the bunch
How in the fuck?

>clearly worse VRM
Like he said, you live in 2016, Maxwell and Pascal GPU are not like Kepler. The reference/FE card are god tier. Overckocking is dead.

>no source
techpowerup.com/reviews/Zotac/GeForce_GTX_1080_Mini/32.html
techpowerup.com/reviews/MSI/GTX_1070_Quick_Silver_OC/32.html
techpowerup.com/reviews/MSI/GTX_1080_Ti_Gaming_X_Trio/33.html

And the chenese AIBs don't use less VRM than the reference

Attached: Screenshot_20180806-105720.jpg (880x684, 216K)

With Pascal, the voltage is locked. Get the cheaper card, you don't need 666666 phases

Heh, when I posted I was literally expecting him to call less VRMs/PCIe connectors, so I'd post the reference PCB, and turns out someone eles did it for me.

Doesn't that mean no OC whatsoever? Just the card auto boosting on its own? This is BS.

Sadly this. No more oc fun with Nvidia, only get the cheaper card with the best cooling. Pic related is now useless with Pascal but we don't know if real overckocking will be back with Amper/volta/turing

Attached: front.jpg (1237x600, 257K)

This. I've been seriously considering for the next upgrade to just grab the cheapest card and slap a Morpheus on it if this keeps up. I have the case space for it anyway.

You're not going to much higher than 2000~2100mhz no matter how much you spend on the damn card for regular use.

It's not a component or VRM issue, it's the peak of the architecture.

Overclocking (on air or low end water cooling) has been dead since clocks stopped scaling with lithography.
Unless a company intentionally releases a clock-gimped, overclockable product it's never worth it to OC modern products beyond their part specification.

For example,
Most 1050Ti ship with 8Ghz capable VRAM, but are factory set at 7000. This is the only case where OCing makes sense and doesnt destroy your product.

JUST WAIT NAVI

BLOWN
THE
FUCK
OUT

Attached: 1532671370631.jpg (500x425, 37K)

is that a footprint for type C i see?

Wat's that. On the top, right to the SLI connector?

They think is a next gen sli

>TFW there are retards here who will be annoyed of it because they can't connect the 1180 Ti to their CRT monitors

No, reference cards are always shit. Go with a dual or triple fan at least. Unless you want to have airport atmosphere in your living room whenever your GPU runs.

>no open access to nvenc
Nice FUD amd cuck.

Attached: wut.png (960x531, 150K)

You still don't have a real product faggot

codeofhonor.com/blog/whose-bug-is-this-anyway
Read the section of this article titled "Your computer is broken".

You don't know shit.

devtalk.nvidia.com/default/topic/1025542/general-discussion/nvfbc-on-geforce/
I'm referring to the chip complex w/ nvfbc and nvifr not just the encode/decode functionality. They used to allow access to it before Pascal and then restricted it when they found out they could nigger people with geforce experience and charge enterprise premiums for access via their pro cards. They locked consumers out of the api and sdk in an update and then told various groups like OBS software that does screen recording to fuck off.

Attached: NVEncSlide.jpg (2000x1125, 341K)

>codeofhonor.com/blog/whose-bug-is-this-anyway

READ YOUR OWN LINK NIGGER. It was during a time when computer hardware sucked dick.
OR
> UNSTABLE MEMORY OVERCLOCKING
> INADEQUATE POWER SUPPLIES
> OVERHEATING

Not everyone runs on donated hardware.
Bit errors are common for an unstable overclock which is why you don't do it to retard levels and run Memtest [TO CHECK FOR BIT ERRORS].
WHAT KIND OF NIGGER RUNS A SETUP WITH AN UNSTABLE POWER SUPPLY?
WHAT KIND OF NIGGER RUNS A SETUP THAT OVERHEATS REGULARLY

YES nigger, if you run a ghetto ass shitty computer, you will get bit errors.
Most people don't and this isn't an age of IBM PCs so MOST PEOPLE WILL NEVER GET A BIT ERROR

> You don't know shit.
No, you're a poorfag, who doesn't understand shit and who likely runs on a shit tier setup that is susceptible to bit errors. PRO tip nigger, people don't run a computer to the point of overheating, over-straining a psu, or overclocked to unstable levels who care about DATA INTEGRITY

Being this fucking retarded.
ECC memory is a fucking meme and buffer ECC an even bigger meme for poorfag brainlets who spend their day dreaming up computer specs as opposed to running a non-poverty tier modern PC

Attached: 1521054583243.jpg (637x637, 64K)

>and then told various groups like OBS software that does screen recording to fuck off.
Is that why OBS NVENC is low performant but not perfect unlike Shadowplay?

Yes that's exactly why. For screen-capture, If you aren't given access to the frame buffer directly on the video card after it renders a frame then you have to export the frame involving the cpu/sys memory then reimport it back to the video card for encoding. The desktop consumer video cards used to allow people to do this during Maxwell then Nvidia's business dev group figured out a way how to nigger more profit from it so they roped off access and only gave consumers access to enc/decode from that point on after a particular maxwell update and disabled it all together for pascal. So, yes.. you have a capture card essentially in your Geforce card that Nvidia wont let you access because they want to route you through Geforce Experience and collect telemetry data off of you for profit. They allow unrestricted access to it for pro/enterprise cards. Hilariously however, when you use Nvidia's cloud GPU service and are using consumer cards, you're essentially using NVfbc/nvifr. They allow access to it on their side through their cloud but don't allow you access as a consumer because they're profit hungry niggers. OBS begged the niggers and they wouldn't give them access. Steam has access to it because they partnered with their nigger tier cloud services.
> NGREEDIA
So, based on shit like this, it's almost gauranteed they're going to nerf/restrict access to any serious hardware changes in their new GPU lineup.

Holy shit, that's all I can say. Thanks for the informative read user, had no idea. Guess that explains why OBS is doing excellent for me, seeing I got a pretty good CPU so NVENC encoding hardly hits performance on my end except certain games that are more CPU hungry.

>They allow access to it on their side through their cloud
Aren't their cloud servers literally using Quadro cards tho? Or did I misunderstood the point you were trying to make here?

He's right, you shouldn't pre-order hardware with review embargo,and it is ineed niche

I'm not interested in 1xnm grandpa shit

The flow is as such :
Out of GPU pipeline -> Frame buffer
Frame buffer can pipe to display and/or to nvenc and then out from there.

With Ngreedia's nerf, you get :
Out of GPU pipeline -> Frame buffer -> CPU -> Sys memory -> NVenc -> sys memory

CPU doesn't get hit that hard but latency/performance of the operation does. You have to hits to the system memory round trip.

W/o the nerf you'd have :
Out of GPU pipeline -> Frame buffer -> Nvenc -> sys memory
Much less work for OBS/OS/CPU/Sys memory. Much less latency. It's essentially behaving as a capture card here.

People started using it as such and taught Nvidia a lesson about a market/feature they didn't perceive. Instead of rewarding their customers with software support, Nvidia tied it off, created a new cloud service, nerfed it for that existing line of GPUs and future, and pushed people to quadros/nvidia's cloud service.

Nvidia's cloud gaming works on a tech known as Grid.
See pic.. It's a slide that shows how it works. Notice anything?
It's the same architecture from It's called Geforce Grid. It's essentially the same GPU sitting in your computer except w/o the software gimp.

Yes, nvidia are ultimately niggers that setup a cloud service to lease you some shit that was present on your own dam GPU sitting in your case

Attached: ngreedia_grid_same_crap_on_geforce.jpg (1163x602, 127K)

onlivespot.com/2012/05/nvidia-unveiled-geforce-grid-cloud.html

Extended detail in case you wanted a complete read and were interested as to how far back this goes.

They locked the feature out in this driver update :
geeks3d.com/forums/index.php?topic=3421.0

and this shows people talking about it being it accessible option before they locked down steam and the software stack :

steamcommunity.com/groups/homestream/discussions/0/490125103626704042/

NGreedia folks.
So, for the upcoming GPU, I expect them to :
> Have new hardware for real time ray tracing that will be nerfed to shit
> Possible Nvlink that will be nerfed to shit
> NVIFR/NVFBC still nerfed

If I don't get these features, i'm flipping my purchase for a 50% ROI on ebay immediately.

>THANK YOU BASED NVIDIA

Attached: 1498106037558.jpg (680x680, 61K)

The correct term is registered and unregistered idiot.
0.02 shekels has been deposited to your account.

Attached: Threadripper-Registered-ECC.jpg (1147x548, 54K)

They still can, an active VGA adapter will work, I have one that does.

So do we believe the typical angled grainy picture or the leaked board PCB?

One is fake, because the leaked PCB is 6+8, this is just an 8

Absolutely unneeded for a consumer GTX card. That would drive up the price on 100% of the cards to please the 0.001% of idiots trying to do tensor workflows on a fucking gaming oriented card.

>it's almost gauranteed

Translation: I pulled this out of my ass and am pretending it's reality because I am and AMD shill and fud is all I know

they only run at 60hz and are prob like 50ms lag.

not sure that there is any adaptors that work at 120hz and have less than 20ms lag which is totally unacceptable.

CRT is stuck on 980 for now or 980ti/TitanXMaxwell if your a rich bitch.