ASRock has unveiled the Radeon AI PRO R9700 Creator graphics card, a new entrant designed specifically for edge artificial intelligence acceleration and professional visualization. The graphics card is built on AMD’s recently released PRO R9700 GPU, utilizing the advanced 4 nm ´Navi 48´ silicon. ASRock puts the chip to full use, enabling all 64 compute units, resulting in 4,096 stream processors, 128 dedicated artificial intelligence accelerators, and 64 ray tracing accelerators. The R9700 distinguishes itself from gaming-focused cards like the RX 9070 XT by doubling the memory to a substantial 32 GB, making it well-suited for complex and large artificial intelligence model workloads.
Optimized for multi-GPU workstations, the Radeon AI PRO R9700 Creator incorporates PCI-Express Gen 5 memory pooling, allowing up to four cards to operate together and efficiently share resources. ASRock has engineered the card for compact professional systems by adopting a true 2-slot form factor, crucial for dense workstation builds. The lateral blower-based cooling solution features airflow cutouts and a recessed backplate to enhance cooling performance, especially in multi-GPU arrangements where thermal management is vital.
Another notable design aspect is the placement of the 16-pin 12V-2×6 power connector at the card´s tail end. This layout makes the R9700 uniquely capable of fitting into 3U rackmount chassis, avoiding the cable bending limitations of top-oriented connectors and improving compatibility with enterprise installations. Measuring 27.1 cm in length and 11.1 cm in height, the card supports four DisplayPort 2.1a outputs, catering to advanced visualization needs. With these features, ASRock extends its professional hardware portfolio, following the Radeon RX 9070 XT Taichi, and aims to meet the rigorous requirements of artificial intelligence and visualization professionals.
