AMD’s Radeon Pro VII Professional Graphics Card Leaks Out – 16 GB HBM2, 7nm Vega GPU, Reportedly Faster Than The Quadro RTX 5000
AMD's Radeon Pro VII graphics card which aims to tackle the professional market segment has leaked out by Videocardz. The graphics card seems to be a PRO variant of the Radeon VII that launched as an enthusiast gaming option for 2019 but the card got discontinued the same year.
AMD's Radeon Pro VII Brings 7nm Vega To The Professional Segment - 16 GB HBM2 Memory & 250W TDP
The AMD Radeon VII was designed to compete against the NVIDIA GeForce RTX 2080 although the GeForce RTX 2080 Ti was already out as green team's flagship offering. The card managed to offer a good improvement over the Radeon RX Vega 64 but ultimately fell to its despite owing to a power hungry chip design and high-temperatures. The card being inefficient against NVIDIA's RTX graphics cards despite using a 7nm architecture meant that it didn't saw a great response from the gaming community as well as AMD had expected.
But the underlying Vega architecture, since its inception, has been known to be a great GPU for professional and datacenter workloads. That is why the upcoming cDNA architecture based GPU design will still feature the fundamentals of Vega at its heart. Based on that, AMD is bringing out its latest professional graphics card, the Radeon Pro VII.
Just like the name suggests, the card will make use of an iteration of the 7nm Vega GPU. We can't say for sure whether this will be the same SKU as the one featured on the Radeon VII but if we go by the naming scheme, it definitely should. In that case, we can expect 3840 stream processors in the Vega 20 GPU. Clock speeds for the variant aren't listed but we get the exact compute power of the card which is rated at 13.1 TFLOPs (FP32) and 6.5 TFLOPs (FP64). Interestingly, the 13.1 TFLOPs FP32 performance is slightly lower than the Radeon VII which hints at a lower clock speed and that can be attributed to the 250W TDP design of the Radeon VII Pro whereas the Radeon VII features a much higher 295W TDP, resulting in higher clock speeds.
Key capabilities and features of the AMD Radeon Pro VII graphics card include:
- Leading Double Precision Performance – With up to 6.5 TFLOPS (FP64) of double-precision performance for demanding engineering and scientific workloads, the Radeon Pro VII graphics card provides 5.6x the performance-per-dollar versus the competition on the AMD Internal Benchmark for AltairEDEM “Screw Auger” views.
- High-speed Memory – 16GB of HBM2 with 1TB/s memory bandwidth and full ECC capability handles large and complex models and datasets smoothly with low latency.
- AMD Infinity Fabric Link – A high-bandwidth, low-latency connection that allows memory sharing between two AMD Radeon Pro VII GPUs, enabling users to increase project workload size and scale, develop more complex designs and run larger simulations to drive scientific discovery. AMD Infinity Fabric Link delivers up to 5.25x PCIe 3.0 x16 bandwidth with a communication speed of up to 168 GB/s peer-to-peer between GPUs.
- Remote Working – Users can access their physical workstation from virtually anywhere for unhindered productivity with the remote workstation IP built into AMD Radeon Pro Software for Enterprise driver.
- High-bandwidth PCIe 4.0 Support – PCIe 4.0 delivers double the bandwidth of PCIe 3.0 to enable smooth performance for 8K, multichannel image interaction.
- Frame Lock/Genlock – Enables precise synchronized output for display walls, digital signage, and other visual displays (AMD FirePro S400 synchronization module required).
- High-resolution/Multi-display Support – Supports up to 6x synchronized display panels, full HDR, and 8K screen resolution (single display) combined with ultra-fast encode and decode support for enhanced multi-stream workflows.
- Professional Application Certification – Optimized and certified with leading professional applications for stability and reliability. The list of Radeon Pro Software-certified ISV applications can be found here.
- ROCm Open Ecosystem – Open software platform for accelerated compute provides an easy GPU programming model with support for O
The FP64 performance is almost 2x over the Radeon VII which is mostly due to the 1/4 rate on the gaming variant whereas the Pro variant comes with 1/2 rate FP64 compute. The card also features 16 GB of HBM2 memory delivering a peak bandwidth of 1 TB/s. There's also a new Infinity Fabric link on the card which is rated at 168 GB/s interconnect while the cards themselves are PCIe Gen 4 compliant.
AMD Radeon Pro Workstation Graphics Lineup:
|Graphics Card Name||Radeon Pro WX 7100||Radeon Pro WX 8200||Radeon Pro WX 9100||Radeon Pro W5700||Radeon Pro W5700X||Radeon Pro VII|
|GPU||Polaris 10||Vega 10||Vega 10||Navi 10||Navi 10||Vega 20|
|Clock Speed (Peak)||1243 MHz||1500 MHz||1500 MHz||1930 MHz||~1850 MHz||TBD|
|Compute Rate (FP32)||5.7 TFLOPs||10.8 TFLOPs||12.3 TFLOPs||8.89 TFLOPs||9.5 TFLOPs||13.1 TFLOPs (FP32)
6.5 TFLOPs (FP64)
|VRAM||8 GB GDDR5||8 GB HBM2||16 GB HBM2||8 GB GDDR6||16 GB GDDR6||16 GB HBM2|
|Memory Bandwidth||224 Gbps||484 Gbps||512 Gbps||448 Gbps||448 Gbps||1024 Gbps|
|Price||$799 US||$999 US||$2199 US||$799 US||$999 US||$1899 US|
Quarto RTX 5000 is up to 68% slower than Radeon Pro VII in "HOPPER DISCHARGE".
It's a benchmark.
— VideoCardz.com (@VideoCardz) May 13, 2020
According to Videocardz, the NVIDIA Quadro RTX 5000 is the main competitor of the Radeon VII Pro and is up to 68% slower in a specific benchmark versus the Vega 20 Pro card. NVIDIA's Quadro RTX GPUs also rely on the older PCIe Gen 3 interface whereas the industry is moving towards PCIe Gen 4. NVIDIA is expected to unveil its own PCIe Gen 4 based platform with its Ampere GPUs for HPC & professional segments tomorrow so we can expect more details.
AMD has been providing 2nd Gen and 7nm based Vega GPU in its Pro lineup for a while now but they have been kept exclusive to Apple's Mac Pro systems. With the Radeon VII Pro, a wider variety of professionals will have access to the GPU. As for now, the Radeon VII Pro is looking like a great card for professionals and in terms of pricing, it is expected to retail at $1899.99 US which is a low price compared to its main competitor while offering better compute horsepower.