Why did both Nvidia and AMD stop making dual GPU cards?

Because crossfire and sli are the biggest pain in the ass to work with.

>shit support
>power usage
>Heat

Always buy best single gpu solution.

because it further accentuates the issues inherent to dual graphics and introduces some new shortcomings

https://www.amazon.com/XFX-Professional-Workstation-Enthusiast-Graphics/dp/B01EFR43ZG

wtf are you talking about

shit's only a year old, dumbass

my old 4870x2 still has higher power usage than just about everything...

>>59227427
>>>/v/

>>59227448
You literally said nothing with that post.

Neither Nvidia nor AMD ever stopped selling multiGPU cards.
It is in fact AMD's long term goal to have multiGPUs become standard analogous to how multicore CPUs became standard.

>>59227499
source?

>>59227517
Some interview with Raja whatshisname.

Gpu pretty much parallel to begin with, dual gpu is oxymoron

>>59227497
>literally
i got triggered.
8/10 troll

>>59227497
You just lack the knowledge required to properly understand the post in question.

>>59227408
AMDs entire plan is to make dual, quad, etc GPUs in Navi, without Crossfire. Idea they want to do is MCM them together.

Microstuttering and driver support were issues that never really got solved. Plus the single chips are so good now that there's no need.

File: bitching fast 3d graphics.jpg (218KB, 800x1051px) Image search: [Google]

218KB, 800x1051px

>>59227408

>>59227427
>Because crossfire and sli
We use split frame rendering with dx12/vulkan these days kiddo B)

>>59228548
I almost had this as separate cards... a cirrus logic for dos, the matrox g400 (got burned on t e mystique) for Windows gui acceleration and then a voodoo for quake... those were the days

After owning a gtx295 back in the day I'm glad they stopped.

>>59229122
lots of old CAD/OGL accelerators usually had a shitty cirrus logic controller to handle 2D until that function was integrated onto the main core

>>59227481
>that
>professional
>workstation
Who do they think they're fooling with that?

>>59227408
Because they're shit. About to throw away my 295x2 because it's literal garbage and I wouldn't be able to sleep with a clear conscience if I actually tricked someone into paying for this piece of shit.

Can't wait to grab myself a 1080 Ti. Don't ever make the same mistake I did and fall for the multi-gpu meme.

>>59227408
When you look at a dual-CPU motherboard, there are dedicated RAM slots for CPU1, and for CPU2; same for a 4-CPU mobo.

The problem is here physical space vs. RAM. You get far better ratios on multiple PCI-e cards than you do trying to cram it all onto one. If you extend the card like this frankenboard, now it won't fit in a lot of cases; those it will fit in, it messes up the air flow for every other component.

>>59229718
if I knew you IRL I would unironically ask you to give it to me because I like housefires of bygone times

I really want a bitchin' dual-chip Quadro to fit in my loaded precision T7400 even if in practice they were limp dick as fuck.

>>59227499

it's everyones long term goal. $/mm2 is going up and will eventually reach a point where fabbing 450mm2-600mm2 monster dies is going to be unprofitable.