Nvidia and Advanced Micro Devices are ramping up the competition in the professional graphics market with new GPUs and other technologies aimed at workstation users running such workloads as virtual reality and high-end gaming.
The graphics card vendors introduced their new offerings this week at the SIGGRAPH Conference in Anaheim, Calif., with officials pointing to a broad array of emerging use cases for the products, from virtual reality (VR) and animations to gaming, complex modeling and simulations. The offerings also are nods to the growing demand within the industry for open-source technologies, according to Raja Koduri, senior vice president and chief architect at AMD’s Radeon Technologies Group.
“We have taken a long, hard look at the workstation space and chose to focus on what we define as ‘the art of the impossible,'” Koduri said in a statement. “Radeon Pro represents a powerful alternative to the old way of doing things, setting aside proprietary solutions for open source ones, and closed ecosystems in favor of greater choice and flexibility.”
At the show, AMD unveiled its new Radeon Pro WX Series of professional graphics cards based on the chip maker’s new Polaris architecture. AMD introduced the Polaris architecture at the 2016 Consumer Electronics Show in January, basing Polaris on a 14-nanometer 3D FinFET transistor design for improved performance and power efficiency. The architecture will more than double the performance-per-watt of previous Radeon graphics technologies, officials said. The company earlier this year rolled out the Radeon RX GPU portfolio based on Polaris.
Within the new Radeon Pro WX Series is the WX 7100 GPU, which is aimed at design engineering and media and entertainment workflows, and professional VR content creation. The WX 5100 can handle design visualization workloads for product development, while the WX 4100 is a half-height design for small form-factor workstations.
In other moves, AMD also is open-sourcing its rendering engine, Radeon ProRender, which will give developers access to the code. It’s part of the company’s expansion of its open-source efforts under the GPUOpen program launched last year by AMD’s Radeon Technology Group to give developers greater access to its graphics technologies. The chip maker also has made its Radeon Rays ray tracing intersection library for GPUs, CPU and accelerated-processing units (APUs) into GPUOpen.com.
AMD officials have made GPUs a key part of the plan to grow revenues and to compete with Nvidia in such areas as gaming, immersive computing, VR and augmented reality (AR).
During a conference call to discuss the latest quarterly financial numbers, AMD CEO Lisa Su noted that sales of GPUs during the first half of 2016 had increased by more than double-digit percentages over the same period in 2015. She also noted graphics cards coming in the third quarter, including the RX 460 and 470.
Nvidia, AMD Roll Out Powerful GPUs for Workstations
“We are feeling very good about our Polaris launch,” Su said during the call, according to a transcript on Seeking Alpha. “We laid out a strategy where we were going after the mainstream, and trying to create really a new experience in terms of both capability and price point, and I think we’ve done that. … Our aspirations in GPU are to certainly have very competitive products across the entire product line and so we’ve talked about working on Vega, which is the next-generation high-end architecture, but in terms of our competitiveness, we’ve executed what we thought we were going to execute and it seems like from both customer reviews and analyst reviews, that it’s pretty well received by market.”
For its part, Nvidia at the SIGGRAPH show rolled out the new Quadro Pascal platform with the Quadro P6000, which comes with 3,840 cores and 12 teraflops of compute performance, which means developers will be able to handle complex designs twice as fast as before, according to officials.
The company continues to ramp up the performance of its GPUs. The launch of the Quadro P6000 comes less than a week after Nvidia announced the $1,200 Pascal-based Titan X GPU, which comes with 12GB of GDDR5X memory and 3,584 cores, and a performance of 11 teraflops. The Titan X’s performance ran by the 8.9 teraflops of compute power offered in the GeForce GTX 1080, which the company unveiled in May.
Along with the Quadro P6000, Nvidia also brought GPU acceleration technology to mental ray, a film-quality rendering product, and introduced the Optix 4, the latest version of the company’s GPU ray tracing engine. The Optix 4 debuted on Nvidia’s DGX-1 supercomputer for such use cases as deep learning and artificial intelligence, which was introduced in April. With the Optix 4, developers can create the fastest interactive rendering for film-size scenes up to 64GB, officials said.