NVIDIA Unveils the Quadro M6000 24GB Graphics Card

[btarunr]

NVIDIA announced the Quadro M6000, its new high-end workstation single-GPU graphics card. Based on the GM200 silicon, and leveraging the "Maxwell" GPU architecture, the M6000 maxes out all the hardware features of the chip, featuring 3,072 CUDA cores, 192 TMUs, 96 ROPs, and a 384-bit wide GDDR5 memory interface, holding 24 GB of memory, double that of the GeForce GTX TITAN X. Its peak single-precision floating point performance is rated at 7 TFLOP/s.

Where the M6000 differs from its the GTX TITAN X is its workstation-grade features. It drops the HDMI 2.0 connector for a total of four DisplayPort 1.2 connectors, supporting a total of four 4K Ultra HD displays. The dual-link DVI connector stays on. There's also an optional stereoscopic 3D connector. The nView MultiDisplay tech provides more flexible display-head configurations than the ones you find on NVIDIA's consumer graphics GPUs; you also get NVIDIA GPUDirect support, which gives better memory sharing access for multi-GPU systems. The M6000 supports most modern 3D APIs, such as DirectX 12, OpenGL 4.5, and Vulkan; with compute capabilities over CUDA, OpenCL, and DirectCompute. NVIDIA didn't reveal pricing.



View at TechPowerUp Main Site

 

[荷兰大母猪]

哦哟好屌哦……然并卵

 

[peche]

Nice skin ... better than the silver one,

 

[Ikaruga]

One lung, a kidney and a forearm.

 

[awesomesauce]

哦哟好屌哦……然并卵

google translate

Oh good yo ...... oh Diao and then the eggs

 

[qubit]

It's gonna cost a small fortune with all that RAM and it doesn't even do compute as well as the last generation, which is actually needed for this type of card.

If I was in the market for this type of product, I'd wait for Pascal for sure.

 

[Darksword]

Pricing will be in quantities of gold bars.

 

[nickbaldwin86]

2 and 6 4K monitors

 

[DeSanta]

Isn't this a bit late for a high end quadro? Titan X has been on the market for a year now.

For a sec. I thought it was based on the new Pascal architecture lol.

 

[peche]

Isn't this a bit late for a high end quadro? Titan X has been on the market for a year now.

For a sec. I thought it was based on the new Pascal architecture lol.

dont care... i want that coler scheme on next reference video cards...

 

[FYFI13]

dont care... i want that coler scheme on next reference video cards...

I don't even care about color schemes anymore. I just want next-gen cards to be at least 30 times cheaper than this massive misunderstanding.

 

[peche]

I don't even care about color schemes anymore. I just want next-gen cards to be at least 30 times cheaper than this massive misunderstanding.

dont think so ....

 

[荷兰大母猪]

google translate

Oh good yo ...... oh Diao and then the eggs

LOL u need to learn modern Chinese

 

[R-T-B]

google translate

Oh good yo ...... oh Diao and then the eggs

I wonder how many eggs it would take to buy this? I mean, frogs lay lots of eggs, so...

 

[NorthboundOcclusive]

but will it run Crysis?

 

[Prima.Vera]

LOL u need to learn modern Chinese

Off topic, and just to let you know we are not all ignorant , there are no such things as Chinese language, Indian language, Phillipino language, etc. There is only Mandarin, Cantonese, Wu, Min, Hindi, Punjabi, Tagalog, etc...

On topic, I was alway wonder what apps are those that can use such huge ammount of VRAM? I used CATIA, SolidWorks, iDEAS a lot, but on memory usage, even on complex assemblies, it didn't go that high...

 

[Caring1]

Isn't this a bit late for a high end quadro? Titan X has been on the market for a year now.

For a sec. I thought it was based on the new Pascal architecture lol.

The M6000 12Gb has been on the market for around the same time as the Titan X.
It seems they have only doubled the amount of Vram, hopefully that won't see the price double as the little brother is still close to $5,000 U.S.

 

[P4-630]

https://translate.google.com/transl...&utm_campaign=hardwareinfo&edit-text=&act=url

"The Quadro M6000 24GB will cost $ 5,000. In Europe, the professional Nvidia cards are sold by PNY."

 

[Serpent of Darkness]

Isn't this a bit late for a high end quadro? Titan X has been on the market for a year now.

For a sec. I thought it was based on the new Pascal architecture lol.

This is pre-Pascal. NVidia is still milking the Maxwell arch until they release a Pascal version, but the popularity of the new arch will be dependent on Pascal on the Gaming side. If their version of HBM2 or GDDRX5 improves performance, plus you have to include the size shrink from 22 to 14nm, they will invest in an upcoming Pascal Quandro and Tesla Card.

On topic, I was alway wonder what apps are those that can use such huge ammount of VRAM? I used CATIA, SolidWorks, iDEAS a lot, but on memory usage, even on complex assemblies, it didn't go that high...

If I am not mistaken, a lot of your renderings (video reproduction and animation type) basically copy framebuffers from CPU to GPU when rendering frames, on the GPU. So when you increase subdivisions or do some kind of increase in AA or adaptive rendering, CPU framebuffer goes up in usage, and so GPU sort of just takes it in in chunks during the processing. A higher framebuffer on the GPU will take it in bigger chunks of information, thus allowing it to complete the frames even faster. There's more space to hold information on the GPU side versus a graphic card with a smaller framebuffer. When the GPU is rendering frames, the CPU still has to hold the information during the rendering process. The usage could go up in a static or dynamic way depending on settings, and if you lock in Framebuffer usages for certain rendering programs and addons.

The only other thing that I can see VRAM going up is if you can render particle effects to the GPU. Still, you'd still be holding information to the CPU framebuffer, but you could probably crank up the resolution/particle amount for the scene.

I see Quadro 6000M v1.0 going down to 2.5k dollars, and v2.0 going up to 5.0K dollars. 6000M is a contender to the AMD Firepro9100, and its' retail price on Newegg is under 3.0k dollars. Only good I see from using 6000M v1.0 and v2.0 cards are the HyperQ and Dynamic Parallelism... AMD will probably respond to a new Polaris Firepro refresh soon too... Following after that, NVidia will probably have a Pascal version Quadro with 12 GBs VRam at best for another 5.0k Dollars. By then, 6000M v2.0 will either sink to 4.9k Dollars or drop somewhere in between Pascal Quadro and 6000M v1.0 price range. Pascal Quadro will probably have a better 64bit Floating Point Precision which will make it more ideal of a workstation card.

 

[Ubersonic]

Isn't this a bit late for a high end quadro? Titan X has been on the market for a year now.

For a sec. I thought it was based on the new Pascal architecture lol.

It's not for gaming, also this is actually a 24GB version of an existing 12GB card so I would assume users have been having VRAM limitation issues prompting Nvidia to launch a bigger one to take advantage of the demand.

 

[Relayer]

Because AMD has 16GB they need moar!

 

[laszlo]

but will it run Crysis?

if won't run no problem; i think super mario bros or wofenstein 3d will use all hardware 100%!

 

[Devon68]

It's a 5000$ card. No normal user would even consider this, only the professionals that need it.
I like the color scheme.

 

[PP Mguire]

I really don't see the point. We have 2 of the 12GB cards that are in a dedicated server for offloading workloads from 2 different teams and I've never seen the VRAM cap. Curious to what professional workload requires this kind of VRAM onboard.