A-V80-P64G-PQ-G

Accelerator cards

Accelerate your memory-intensive workloads with the AMD Alveo? V80 accelerator card. This flagship compute accelerator is architected for memory-bound applications that demand FPGA hardware adaptability to handle custom data types and custom data movement while scaling for large data sets at network line rates. Featuring a powerful AMD Versal? HBM adaptive SoC, the card integrates HBM2e DRAM to overcome memory bottlenecks common in compute-intensive applications. As a network-attached accelerator, the card features 4x200G connectivity for high-speed packet processing and compute clustering—all in a full-height, ? length dual-slot form factor. Scale your infrastructure to hundreds of accelerators over Ethernet or leverage the card’s MCIO cable for clustering over PCIe?. As a platform built for traditional FPGA developers, the Alveo V80 accelerator card is paired with the Alveo Versal Example Design (AVED) for ease of bring up using the AMD Vivado? Design Suite.

特色說明

HARDWARE ADAPTABLE FOR MEMORY-BOUND WORKLOADS

  • Massive parallelism and hardware adaptability for workload optimization
  • HBM2e for memory-bound applications with large data sets

2X PERFORMANCE VS. ALVEO U55C1 COMPUTE ACCELERATOR

  • 2X logic density and memory bandwidth
  • 4X network bandwidth for packet processing and large-scale compute clustering

FAMILIAR DEVELOPMENT FLOWS FOR FPGA DESIGNERS

  • Alveo Versal Example Design (AVED) for ease of hardware bring-up
  • Support for Vivado Design Suite and traditional RTL flows

資源下載

下一階段的系統
歡迎與我們討論系統開發需求