There’s a new kid on the block and its name is Pascal!
You don’t want to waste time worrying whether your computing power can handle the vast geometry of a 3D model or if it can render that ridiculously complex simulation or hyper-realistic VR experience. Lucky for you, there’s a technology manufacturer who enjoys resolving performance challenges just like these, putting Moore’s Law into practice every step of the way. And with that in mind I’m delighted to introduce to you NVIDIA’s most recent addition to the Quadro family - NVIDIA Pascal.
With growing pressure being put on production teams to deliver content with higher fidelity and greater file sizes, it’s important that technology manufacturers take on the challenge of providing solutions that allow our designers and VFX wizards to achieve more - constantly pushing the boundaries of what is possible. Earlier this year for example, NVIDIA launched the very impressive Quadro M6000. This graphics card was already blowing our socks off, armed with 24GB of GDDR5 memory and 3,072 CUDA cores. And with support for multiple 4K screens, the M6000 was comfortable with even the most challenging workloads.
NVIDIA weren’t content at stopping there - and they aren’t one for sitting on a product by the looks of it. The Pascal Quadro range is hot off the production line and it is now possible to purchase the P5000 and P6000 - with the P6000 being the most powerful graphics card on the market.
The architecture for NVIDIA Pascal is designed around five technological progressions which NVIDIA beautifully describe as having “an infinite appetite for computing”. Scalability, power and speed is the aim of the game.
The Pascal GPU is the world's largest FinFET chip ever built. With 150 billion transistors built on bleeding-edge 16 nanometer FinFET fabrication technology, you’re getting supercharged performance that can handle near-infinite computing needs.
Delivering over 5 TeraFLOPS of double precision performance for HPC workloads, the compute architecture inside this GPU is insane and the most powerful ever built.
NVIDIA have developed a high-speed bidirectional interconnect - trademarked as NVIDIA NVLink. Designed to scale performance across multiple GPUs, you can look forward to greater speed in interconnect bandwidth.
Promoting unrivalled compute efficiency, CoWoS® (Chip-on-Wafer-on-Substrate) with HBM2 is NVIDIA’s innovative approach to memory design - offering a mega boost in memory performance. Time to trade in your NVIDIA Maxwell architecture for a more powerful model.
And for those of you who geek out to neural network training, Pascal lets AI algorithms deliver real-time responsiveness for deep learning inference. 16-bit floating point instructions deliver over 21 TeraFLOPS.
This is a very exciting piece of kit and we’re already seeing great demand for it. For more information on how this new technology can enahnce your current pipeline - don’t hesitate to get in touch.