AMD Unveils FirePro W9100

by Hilbert Hagedoorn on: 03/27/2014 08:43 AM | source: | 8 comment(s)

Advanced Micro Devices has launched their new flagship professional graphics card, the Firepro W9100. The latest addition to the Firepro family of professional graphics cards is an impressively powerful unit designed for emerging demands in computation and content creation. The Hawaii based workstation graphics card features an industry-first 16Gb of GDDR5 video memory with a 320 Gb per second memory bandwidth.

Through the W9100 AMD set out to provide the professional realms of content creation and computation with an industry-leading card in compute performance. Featuring 2.67 gigaflops of double-precision compute performance with over 5 gigaflops of single-precision compute performance, the W9100 is the first professional graphics solution to break 2 teraflops in double-precision compute performance.

This is the full Hawaii GPU being used, and that means 2816 Stream processors, 176 TMUs and 64 ROPs. The W9100 has also be engineered with AMD's GCN architecture and features DirectGMA technology streamlining the data flow between the GPU and SDI I/O board. For video content creation professionals the card will drastically reduce encoding time while utilizing OpenCL for real-time 4k rendering.

"It's powered by OpenCL and it's really designed and optimized for real-time 4k" said David Cummings, Senior Director and General Manager of AMD Professional Graphics.

The Firepro W9100 is not yet available for purchase, today simply marks the official launch of the product line. AMD has yet to release other key technical specifications for the new flagship professional graphics card such as the core and memory clock speeds along with the memory bus width.

2 pages 1 2

Cryio
Senior Member

Posts: 556
Joined: 2011-09-01

#4789800 Posted on: 03/27/2014 09:13 AM

The W9000 was already mopping the floor with the K6000 as it is. This is just showing off

)

One thing I don't understand. If W9000 based on the 7970 was so fast, how is that the 290X based W91000 is just slightly faster?

Fox2232
Senior Member

Posts: 11808
Joined: 2012-07-20

#4789805 Posted on: 03/27/2014 09:49 AM

Thermal limitation. There are nearly no changes in power efficiency, therefore they are limited by thermal limits and 16GB eats quite some energy too.

BLEH!
Senior Member

Posts: 6365
Joined: 2010-10-17

#4789907 Posted on: 03/27/2014 03:37 PM

That's a LOT of VRam. Good for all those gamers wanting many, many 4K panels

Lane
Senior Member

Posts: 6361
Joined: 2005-02-25

#4790071 Posted on: 03/27/2014 07:55 PM

The W9000 was already mopping the floor with the K6000 as it is. This is just showing off

)

One thing I don't understand. If W9000 based on the 7970 was so fast, how is that the 290X based W91000 is just slightly faster?

- W9000 = 4Tflops SP / 1Tflops DP..
- W9100 = ~ 5.2 Tflops SP / ~ higher than 2.0 Tflops DP .. ( theoricaly max DPFP is 2.6Tflops )....

Going to a 1:2 DP rate maximum is big ( even if sustained we should be at 1:2.5 rate)

You get 2x more DP than on the W9000 and it should be slighty higher of a TitanZ who is made with 2x GPU GK110.

CPC_RedDawn
Senior Member

Posts: 9918
Joined: 2008-01-06

#4790075 Posted on: 03/27/2014 08:00 PM

That's a LOT of VRam. Good for all those gamers wanting many, many 4K panels

I would actually love to see some gaming benchmarks done on this card.

Never ever seen it done before on AMD professional grade GPU's before.

I am sure they wouldn't be much better than current mainstream GPU. As DP has next to no benefit for games.

Still hope this filters down into their mainstream lineup, using the full Hawaii core.

2 pages 1 2

Post New Comment

Click here to post a comment for this news story on the message forum.