Nvidia is finished
http://gpuopen.com/
@op lol yeah right bro nvidia makes gaming processors not gpus dumbass
>>52632802
>such access will also enable easier porting from current-generation consoles (XBox One™ and PlayStation 4) to the PC platform
how mad sony and ms about that? or architecture on consoles belongs to AMD and they can tell anything about it?
>>52632802
wait, are amd cards now no-blobs open source?
that actually does change a lot.
watch Xbone switch to ARM / PowerVR now kek
>>52632836
PS4 and nvidia are the only ones getting shafted.
too bad tho since Sony are the good guys as well.
>>52633097
>sony
>good guys
>>52632813
Actually this is kinda true.
amd makes actual GPUs, nvidia makes DirectX compatible hardware. Throw DirectX out the window and nvidia becomes shit.
>>52633108
Eh, still better than Microsoft.
>>52633114
What about them cuda cores tho
>>52632802
>instead of making better gpus just invest in some apis no one will ever use
just like mantle or whatever it was called
>>52633160
what's that?
>>52633255
fucking wow
It's the reason you pay $10,000 for 2 tesla GPUs to do medical imaging.
GPUs do a lot fucking more than play games kid.
>>52633239
AMD doesn't have to do much with GPUOpen.
I'm fairly certain that companies that make game engines will be more than happy to implement their things directly into the API.
>>52632802
It's too bad that Nvidia still makes the most powerful cards.
We use only nvidia in our clusters at work for scientific computing. This requires us to use proprietary drivers, but AMD simply doesn't make competitive cards in the high-end market (especially outside of consumer-facing use cases).
>>52633294
Nvidia cards aren't more powerful, they 'just' write way better drivers than AMD.
This won't be a problem in the future for AMD with the much more low leven APIs DX12 and Vulkan.
I really like the way AMD is going, and i'd buy AMD over Nvidya any day, just because nvidia is such a shitty company with scummy business practices.
I just hope they don't fuck up Polaris, because if that turns out to be bad then AMD is done.
>>52633097
hm, why exactly? ps4 uses same AMD chip as xbone with slightly different focus on performance
>>52633326
>Nvidia cards aren't more powerful, they 'just' write way better drivers than AMD.
HAHAHAHAHAHA LEL
>>52633280
>maxwell
>compute heavy
lol no
there's a reason why all of nvidias compute stuff relies on Kepler still, maxwell is essensially a dx11 asic
>>52633288
>Introducing GPUOpen — an AMD initiative designed to enable developers to create ground-breaking PC games, computer generated imagery and GPU computing applications for great performance and lifelike experiences using no cost and open development tools and software.
>AMD doesn't have to do much with GPUOpen.
>>52633799
>designed to enable DEVELOPERS (meaning game devs/engine devs, not AMD doing the work)
>using NO COST (meaning AMD doesn't have to pay for development)
are you dense?
>>52633759
You're retarded and clueless faggot
Kepler is only for FP64 compute
For everything else, Nvidia has replaced with Maxwell
http://www.anandtech.com/show/9776/nvidia-announces-tesla-m40-m4-server-cards-data-center-machine-learning
http://www.anandtech.com/show/9574/nvidia-announces-grid-20-tesla-m60-m6-grid-cards
But it shows how ignorant and clueless faggots like you are, opening your big fat mouth and commenting on things you don't understand
>>52633126
So, what you're basically saying is "the enemy of my enemy is my friend"?
>have you tootled your 3D models?
sounds lewd
![61059[1].png](https://i.imgur.com/k2QN3PW.png "61059[1].png")
>>52633239
Mantle broke the ice.
They had a tonne of big name studios signed on to use Mantle, EA being one of the more prominent names among them.
Mantle coming about then forced, or give Microsoft the sufficient technology to move on with DX12, the core of which is supposedly near-identical to that of Mantle.
To say Mantle was a failure is simply naive. It succeeded in it's primary role of reducing CPU load by offloading what it could to the GPU. It had, as I have previously stated, many big name studios signed on to use it in future titles.
While the image I have used is maybe not the best to represent the ideal use-case, it still proves the point that Mantle works.
Should it provide a 10fps boost to CPUs that are in no way lacking in power, then picture what it will do for CPUs that do feel the strain of running a game like Battlefield 4.
>>52633759
Stop posting. You're literally retarded.
You don't even know what CUDA is, and yet you are here trying to tell people that they are wrong.
>>52633126
Yeah no. Sony has a laundry list of things to hate. Microsoft is at least consistent in their blunders.
>>52633402
Amd share holder detected. I reeealy want to buy some shares too, but they just can't catch first place. It's like they want to play second fiddle.
>>52633097
>Sony
>good guys
You mean the company that infected more than half million computers with a rootkit made with stolen GPL code to stop people from copying songs from a CD?
https://www.schneier.com/blog/archives/2005/11/sonys_drm_rootk.html
>>52633097
>Sony are the good guys as well.
If good is leaking credit card info twice since the release of the PS3 and getting hacked by NK, employee information leaking out, and giving in to demands of "cyber-terrorist" then I don't know what good is.
>>52633799
you realize this is make your own API kit?
Developers that make their own engines are smiling right now.
Not to mention the ability to make compute applications that do what you need them to do very close to hardware.
For example proper supercomputers created specifically for narrow set of tasks on hardware level, you can be not like them but closer to them if this thing is what they promise this thing to be.
>>52633097
>too bad tho since Sony are the good guys as well.
https://www.sony.com/be-moved/
>>52633160
Can't access the cuda cores in current gen geforce and titan cards, only quadforce and Tesla, and those 2 still aren't as good as the firepro equivalents
>>52633239
or gameworks except you don't have to pay anyone...
>>52633732
>>52633239
Mantle was the API that inspired both Vulkan and DX12. In fact, a fair amount of the Vulkan specs is pretty much copy-pasted from the Mantle one.
>>52634784
>>52634807
And there you have it. AMD deserves praise for at least trying to push open standards that end up benefiting everyone.
>>52632802
>2006
>AMD announces CTM, a ground-breaking API to unleash the power of GPUs
>2014
>AMD announces Mantle, a ground-breaking API to unleash the power of GPUs
>2016
>AMD announces GPUOpen, a ground-breaking API to unleash the power of GPUs
>2016
>AMD drivers are still garbage
>>52635046
Pretty much this.
All of these initiatives and rewrites and they still put out broken drivers.
AMD saving gaming one open standard at a time
>>52635071
You're saying that like nVidia's drivers are any less broken.
They only put out a somewhat decent driver every 3 months. Everything in between is a broken mess titled "GameReady" that breaks everything.
How many AMD powered consoles are there compared to nvidia cards being used?
>>52635326
AMD hardware is in all of them (yes, AMD provides the gpu for the wii u).
>>52634131
>didnt rub the edge of his CD with magic marker to avoid the DRM
yes that worked
>>52632802
Vulkan never
