AMD Unveils Radeon Pro W7900

AMD has introduced a new Radeon Pro W7900 GPU, designed specifically for AI workstations. This dual-slot version offers a compact design while maintaining high performance. Priced at $3,499, the GPU aims to deliver top-tier capabilities for professional and creative workflows.

Enhanced AI Performance

The Radeon Pro W7900 features 6,144 shaders and 48GB of ECC GDDR6 memory. This ensures reliable and efficient performance for demanding AI and engineering tasks. The GPU’s compact dual-slot design allows users to install up to four units in a single system, which wasn’t possible with the previous triple-slot model.

Competitive Edge in AI Workloads

AMD claims the Radeon Pro W7900 offers better performance per dollar compared to its competitors. It provides up to 52% better performance per dollar in professional workloads compared to NVIDIA’s RTX 6000 Ada. In AI-specific tasks, the W7900 offers up to 38% better value, making it a cost-effective solution for AI professionals.

The GPU uses the Navi 31 architecture, providing 96 Compute Units, 96 Ray Tracing cores, and 192 AI accelerators. It delivers up to 123 TFLOPs of half-precision performance and 864 GB/s of memory bandwidth. The Radeon Pro W7900 maintains a 295W TDP and requires two 8-pin PCIe power connectors. It features three DisplayPort 2.1 outputs and one mini-DisplayPort 2.1 output.

Alongside the Radeon Pro W7900, AMD will release ROCm 6.1.2, enhancing support for AI frameworks like TensorFlow and providing beta support for Windows WSL-2. This update aims to improve multi-GPU setups and overall AI performance.

The Radeon Pro W7900 is set to hit the market on June 19, offering a powerful and compact solution for AI and professional workloads.