Nvidia has announced new entries in its Quadro workstation product family, including the Quadro GP100 - the first card to bring the company's high-speed NVLink connectivity to the desktop platform.
Sitting between the mainstream GeForce line and accelerator-only Tesla range, Nvidia's Quadro family is aimed squarely at workstation use. Accordingly, you'll find a range of features not usually available on mainstream graphics cards: higher double-precision compute performance, larger amounts of memory, and now a high-speed multi-GPU backbone technology dubbed NVLink and designed to offer improved performance over traditional SLI connectivity.
'Professional workflows are now infused with artificial intelligence, virtual reality and photorealism, creating new challenges for our most demanding users,' claimed Bob Pette, vice president of Professional Visualisation at Nvidia, of the lineup. 'Our new Quadro lineup provides the graphics and compute performance required to address these challenges. And, by unifying compute and design, the Quadro GP100 transforms the average desktop workstation with the power of a supercomputer.'
The Quadro GP100, as hinted at by Pette, represents the jewel in the Quadro crown. The Pascal-based card represents the fastest Quadro Nvidia has yet made, boasting 12 teraflops of 32-bit floating point performance and 5 teraflops at 64-bit precision across an impressive 3,584 CUDA cores. The card also includes 16GB of High Bandwidth Memory 2 (HBM2), and is the first non-server product to feature the company's high-speed NVLink technology in place of the traditional SLI connector for bridging two cards in a single system. This stands in contrast to Nvidia's previous NVLink implementations, which have been used alongside specialist motherboards as a replacement for PCI Express in card-to-host communications.
Elsewhere in the range, the specifications are more sedate: The remainder of the Pascal-based Quadro cards announced by the company lack both HBM2 and NVLink, though still improve upon the performance of their predecessors. Nvidia is particularly keen to point out the potential behind the company's new Quadro P4000 cards in the digital signage market, promising support for up to 32 4K-resolution displays from a single machine by combining eight P4000 cards with two Quadro Sync II cards.
Pricing for the new Quadro cards has yet to be confirmed, with availability scheduled for March. More information is available from Nvidia's official press release.
Sitting between the mainstream GeForce line and accelerator-only Tesla range, Nvidia's Quadro family is aimed squarely at workstation use. Accordingly, you'll find a range of features not usually available on mainstream graphics cards: higher double-precision compute performance, larger amounts of memory, and now a high-speed multi-GPU backbone technology dubbed NVLink and designed to offer improved performance over traditional SLI connectivity.
'Professional workflows are now infused with artificial intelligence, virtual reality and photorealism, creating new challenges for our most demanding users,' claimed Bob Pette, vice president of Professional Visualisation at Nvidia, of the lineup. 'Our new Quadro lineup provides the graphics and compute performance required to address these challenges. And, by unifying compute and design, the Quadro GP100 transforms the average desktop workstation with the power of a supercomputer.'
The Quadro GP100, as hinted at by Pette, represents the jewel in the Quadro crown. The Pascal-based card represents the fastest Quadro Nvidia has yet made, boasting 12 teraflops of 32-bit floating point performance and 5 teraflops at 64-bit precision across an impressive 3,584 CUDA cores. The card also includes 16GB of High Bandwidth Memory 2 (HBM2), and is the first non-server product to feature the company's high-speed NVLink technology in place of the traditional SLI connector for bridging two cards in a single system. This stands in contrast to Nvidia's previous NVLink implementations, which have been used alongside specialist motherboards as a replacement for PCI Express in card-to-host communications.
Elsewhere in the range, the specifications are more sedate: The remainder of the Pascal-based Quadro cards announced by the company lack both HBM2 and NVLink, though still improve upon the performance of their predecessors. Nvidia is particularly keen to point out the potential behind the company's new Quadro P4000 cards in the digital signage market, promising support for up to 32 4K-resolution displays from a single machine by combining eight P4000 cards with two Quadro Sync II cards.
Pricing for the new Quadro cards has yet to be confirmed, with availability scheduled for March. More information is available from Nvidia's official press release.
MSI MPG Velox 100R Chassis Review
October 14 2021 | 15:04
Want to comment? Please log in.