AMD announced a new flagship workstation graphics card it calls the FirePro W9100, at a press conference in California yesterday. This new graphics card brings the advances of the GCN 1.1 Hawaii architecture such as the Asynchronous Compute Engine (ACE) and better performance per watt to professional graphics customers.
While AMD gave us some headline figures about the performance of the new FirePro flagship (above) it left out the key specification details such as clock speeds and memory bus width. However we know the new FirePro shares a lot of the features of the flagship consumer graphics card, the Radeon R9 290X. We also got to see some benchmark results showing performance against leading competitor cards from Nvidia – but of course these will be cherry picked to show the new FirePro in the best possible light.
AMD's FirePro W9100 certainly boasts some great capabilities. The new workstation card is capable of simultaneously driving six displays at 4K resolution. For single precision calculations you can get five teraflops out of this card. Double precision calculations perform at two teraflops. It also blasts the competition away in OpenCL calculations as you can see in the chart below, also offering excellent multi-GPU scaling.
Thanks to the AMD design win and inclusion in the new Mac Pro, Adobe has been busy working to add OpenCL support to its Creative Cloud suite including Premiere Pro. This will be a boon to many content creators on PCs too. AMD execs also said that the 16GB of graphics memory will help improve workflow speed in programs like SolidWorks and Maya. Much of the productivity emphasis was directed at 4K content creators.
The AMD FirePro W9100 is due out in April; pricing will be announced nearer to that time.
