NVIDIA announces 12GB Quadro K6000 "fastest GPU ever"

NVIDIA has announced at Siggraph, the annual computer graphics conference, the Kepler-based Quadro K6000. The company says it's "the fastest and most capable GPU ever built" with twice the graphics capability of the Quadro 6000 predecessor. Onboard is 12GB of the "world's largest and fastest" GDDR5 graphics memory, enabling companies like Nissan to load nearly complete vehicle models, for instance. Other specs include 2,880 streaming multiprocessor cores, four simultaneous displays at up to 4K resolution, ultra-low latency video I/O and large scale visualization support.

NVIDIA also made it clear that the memory capacity involved in the K6000 allows previews of scenes with a much more final vision than has ever been offered before:

It’s going to allow artists to preview their scenes much earlier in the pipeline. In this case, with the K6000, it’s all interactive.”

The company has yet to offer pricing for the new card which will be available from this fall, but if their K5000 is anything to go by at around $1800 (on Amazon) and with the initial launch price at $2,249, you're going to need deep pockets to acquire one of these.

Source: Engadget via NVIDIA | Image: NVIDIA

Report a problem with article
Previous Story

Tim Cook: high-end smartphone market hasn't peaked

Next Story

Leaked images & video claims to show new 5-inch Blackberry A10

22 Comments

View more comments

In for two in SLI. Well not really. How do these compare to say a Titan for gaming? Though Im going to assume these are meant for a different demographic.

JHBrown said,
In for two in SLI. Well not really. How do these compare to say a Titan for gaming? Though Im going to assume these are meant for a different demographic.

They're not, they're for AutoCAD and other GPGPU type tasks, maybe some 3D modeling work and so on. I honestly don't think they'd do good for gaming.

JHBrown said,
In for two in SLI. Well not really. How do these compare to say a Titan for gaming? Though Im going to assume these are meant for a different demographic.

Those cards are built mostly for CAD/CAM and have OpenGL implementations rather than DirectX.

virtorio said,
I've had two (supposedly work) laptops with mobile Quadro graphics and they play games just fine.

They play them just fine, but they're optimized for other tasks and aren't as efficient at gaming as cheaper game cards.

HawkMan said,

They play them just fine, but they're optimized for other tasks and aren't as efficient at gaming as cheaper game cards.

Actually most 3d graphics suites and cad programs today on windows run either DX or let's you choose between DX and OGL. Most will run DX now due to higher performance. Back in DX9 and before artists and designers switched between them more since DX was faster but OGL was more stable. DX has since become as stable as well so...

HawkMan said,

Actually most 3d graphics suites and cad programs today on windows run either DX or let's you choose between DX and OGL. Most will run DX now due to higher performance. Back in DX9 and before artists and designers switched between them more since DX was faster but OGL was more stable. DX has since become as stable as well so...


Still the card is made for design. AutoCAD, Photoshop, video editing etc.
A same priced card that's made for gaming will outperform this one into oblivion.

Worse. Any card in the same price class that's developed for gaming, will easily outperform this one.

If you want this card for gaming, not recommended.

What's the difference between a workstation graphics card and a gaming graphics card? I know workstation cards cost more $$$. However, I've sometimes thought if they're just a re-branded older generation high-end card. I don't expect CAD to be more complex than today's games.

68k said,
What's the difference between a workstation graphics card and a gaming graphics card? I know workstation cards cost more $$$. However, I've sometimes thought if they're just a re-branded older generation high-end card. I don't expect CAD to be more complex than today's games.

Workstation cards has additional instruction sets and often operate at higher accurate, bot of these cause the card to operate slower. Bt the extra instructions as well as the accuracy is important for CAD designers.

Difference is also in shader clock speeds, which are important for gaming.
These cards often come with a lot more RAM to load a whole rendering project.
Performance is aimed at loading an entire rendering project at once rather then frame-by-frame performance.

At least price is becoming more reasonable for these cards. I work with video and 3D titling and picked up a GTX 780, the "cuda cores" are what Adobe suites use for hardware acceleration through their "Mercury" rendering engine. Simply put, I can render HD video along with ray traced 3D graphics at roughly 10x the speed as without it on my i7 3770k, 32GB system. When you're editing a 30 second commercial, you constantly render a few seconds at a time to check how it looks after small changes, this is a huge time saver. These cards typically have many more cuda cores and RAM that I would guess could offer 2-3x the performance of Titan...but I could be wrong, with so many cuda cores on gaming cards these days, the performance gap is narrowing....but the RAM is what gets sucked up so quickly. My 32GB + 64SSD system RAM all gets used when editing heavy stuff. The paltry 3GB Graphics memory is gone after the first few seconds of work.

Commenting is disabled on this article.