AMD Launches The New Radeon Pro Duo At $999
AMD announced the Radeon Pro Duo graphics card targeting the professional consumer market. This new GPU is based on the fourth-generation Graphics Core Next (GCN) architecture and can perform graphic and arithmetic instructions in parallel.
The benefit of the dual GPU Radeon Pro Duo graphics card is that designers can now use two software packages simultaneously. Professionals can take advantage of the parallel computing power of the Radeon Pro Duo graphics card to accelerate compute-intensive tasks in computer-aided design, computer-aided manufacturing, and other applications that support OpenCL.
Even if your application doesn’t support multi-GPU acceleration, this professional graphics card makes it possible to dedicate one GPU to the functions of separate applications concurrently, or to work with multiple 4K video streams in real-time. According to AMD, live content creation using the first GPU, with real-time rendering and/or ray tracing on the second GPU, is now possible.
Ogi Brkic, general manager, professional graphics, Radeon Technologies Group, AMD, said:
Today’s professional workflows continue to increase in complexity, often demanding that creators switch between a wide variety of applications to progress their work, pausing efforts in one application while computing resources are focused on another. We designed the Radeon Pro Duo to eliminate those constraints, empowering professionals to multi-task without compromise, dedicating GPU resources where and how they need them. It’s a continuation of our promise for Radeon Pro: to provide greater choice in how professionals practice their craft, enabling superior multi-tasking, accelerated applications, and powerful solutions for advanced workloads like VR.
The full-height single-slot Radeon Pro Duo features two Polaris 10 cores both operating at 1,243MHz with 2,304 stream processors each. Total compute power is listed at 11.45 TFLOPs. The 32GB of GDDR5 (16GB per core) operates at 7GHz for a total of 448GBps bandwidth. The Radeon Pro Duo graphics card can drive up to four 4K displays at 60Hz or one 5K or 8K display at 60Hz via the three DisplayPort 1.4 HBR3/HDR Ready outputs and one HDMI 2.0 output. Power is supplied via an 8+6-pin power connector.
Planned availability is the end of May at an expected $999 with a 2-year limited warranty and 24/7 technical support.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Header Cell - Column 0 | AMD Radeon Pro Duo |
---|---|
GPU ARCHITECTURE | GCN 4th Generation |
STREAM PROCESSORS | 2,304 x2 |
PEAK PERFORMANCE | 11.45 TFLOPS |
PEAK TRIANGLES | 9.94BTps |
PEAK ENGINE CLOCK | 1,243MHz |
MEMORY TYPE | 32GB GDDR5 (16GB per GPU) |
MEMORY INTERFACE | 256 bit x2 |
MEMORY DATA RATE/ BANDWIDTH | 7Gbps / 448GBps |
TYPICAL BOARD POWER | <250 W |
AMD FREESYNC™ TECHNOLOGY | Yes |
AMD CROSSFIRE™ PRO TECHNOLOGY | Yes |
DISPLAY OUTPUT CONNECTORS | 3x DP, 1x HDMI |
DISPLAYPORT VERSION | 1.4 – HBR3/HDR Ready6 |
5K SUPPORT @ 60 HZ | 1x single-cable 5K monitor, or 2x dual-cable 5K monitors |
8K SUPPORT @ 30 HZ | 1x single-cable 8K monitor |
8K SUPPORT @ 60 HZ | 1x dual-cable 8K monitor |
DISPLAY COLOR DEPTH | 10-bit Support |
STEREO 3D OUTPUT (3-PIN DIN) | No |
FRAMELOCK/GENLOCK SUPPORT | No |
API SUPPORT | DirectX 12 OpenGL 4.5 OpenCL 2.0 Vulkan 1.0 |
OPERATING SYSTEM SUPPORT | Windows 7 64-bit Windows 10 64-bit Linux 64-bit |
FORM FACTOR | Full-Height Single Slot 12” Length |
There's a budget GeForce GPU selling in China that not even Nvidia knew it made — RTX 4010 turns out to be a modified RTX A400 workstation GPU
US to patch loopholes that allow China to buy banned AI GPUs from other countries — new regulations include national quotas on GPU exports and a global licensing system
-
valeman2012
Ashes of Singularity a game that updated to make Intel Kaby Lake CPU or GEFORCE Benches weaker. If you put ATI/AMD Cards or CPU in it thats another story.19606218 said:Yeah, but how does it do with Ashes of Singularity?
-
LouisCyphre You're missing in the specs above what bus is it. It looks like PCie x16, but I want to be sure.Reply -
MikeMIke86 Spoken like a fan boy, NVIDIA has intentionally made several games more difficult for AMD to optimize and you mention one game that's always done better due to AMDs strong async compute ability.Reply
BTW Ryzen is coming for your i7, it's already getting beaten on average (in gaming) with 3600mhz ram and several optimizations.
I'm not a fan boy, I run the cheapest of both sides.. AMD gpu's and Intel i5 cpu's, the 8series processors were a joke but this fresh breath of air is fantastic for competition, kinda nice to see Intel choke after years of mobo pin changes in order to force feed more hardware for a measly 5-10% gain over the previous generation.. so I guess I've began grudging against Intel due to their greed more so than anything.
For all the reasons above I still run a 2500k 4.7gHz and a 3560k 4.5gHz, no reason to upgrade for 5% fps gains or 10% overall gains. -
derekullo 19606848 said:You're missing in the specs above what bus is it. It looks like PCie x16, but I want to be sure.
Were you expecting AGP 8x ?
-
bit_user The full-height single-slot Radeon Pro Duo features two Polaris 10 cores GPUs both operating at 1,243MHz with 2,304 stream processors each. Total compute power is listed at 11.45 TFLOPs. The 32GB of GDDR5 (16GB per core GPU) operates at 7GHz for a total of 448GBps bandwidth.
Please use correct nomenclature. The concept of GPU cores is already confused enough, with the industry using the term to refer to SIMD lanes, rather than what would classically be known as a CPU core. -
Gam3r01 19607484 said:The full-height single-slot Radeon Pro Duo features two Polaris 10 cores GPUs both operating at 1,243MHz with 2,304 stream processors each. Total compute power is listed at 11.45 TFLOPs. The 32GB of GDDR5 (16GB per core GPU) operates at 7GHz for a total of 448GBps bandwidth.
Please use correct nomenclature. The concept of GPU cores is already confused enough, with the industry using the term to refer to SIMD lanes, rather than what would classically be known as a CPU core.
Out of curiosity, what part is incorrect? -
bit_user
See where I used strike-through? Those are my corrections. GPU and CPU refer to the physical chips, each of which may have multiple cores.19607486 said:19607484 said:The full-height single-slot Radeon Pro Duo features two Polaris 10 cores GPUs both operating at 1,243MHz with 2,304 stream processors each. Total compute power is listed at 11.45 TFLOPs. The 32GB of GDDR5 (16GB per core GPU) operates at 7GHz for a total of 448GBps bandwidth.
Please use correct nomenclature. The concept of GPU cores is already confused enough, with the industry using the term to refer to SIMD lanes, rather than what would classically be known as a CPU core.
Out of curiosity, what part is incorrect?
In fact, the Polaris 10 GPUs each have 36 cores, which AMD refers to as "Compute Units". They're really not much different than CPU cores, though.