| Data Interface |
Aurora 64B/66B encoding
- 1 x QSFP-DD input and output, 8 lanes QSFP-DD input and output, 8 lanes
- Each lane supporting up to 12.5 Gb/s line rate (100 Gb/s aggregate line rate)
|
| Host Interface |
- PCIe Gen3 x16 with DMA engine, total maximum theoretical bandwidth of 128 Gbps
|
| Hardware Platform |
- Zynq UltraScale+ MPSoC: up to 1143K logic cells and 522K LUTs in programmable logic
- Quad Arm Cortex-A53 @ 1.5 GHz and dual Cortex-R5 @ 600 MHz
- 8 GB eMMC flash for boot and OS storage, expandable up to 128 GB; micro-SD slot
|
| Cooling |
|
| Form Factor |
- 254 mm x 111.15 mm, 3/4 length, full height, dual-slot PCIe card with active cooling
|
| Measured Performance |
- High-Performance Data Throughput: Achieved 94 Gb/s (11.7 GB/s) card-to-host throughput by transferring Aurora receive data to the host through the DMA engine
- Validated Performance: Sustained the measured throughput over a 12-second test run on a custom-built host PC with Native Flow Control enabled
- Host-Optimized Data Transfer: The card delivers its full measured throughput when the host platform can keep pace. If the host cannot sustain the incoming data rate, throughput falls to the rate the host can accept. With Native Flow Control enabled, the link is paced to the host instead of dropping data
|
| Software Feature |
- Supports high-speed data streaming from the capture card to the Host PC via PCIe
- The Frame Grabber architecture is optimized for low-latency data transfer
- Native Flow Control (NFC) for loss-free capture when the host or disk cannot keep pace
- Multi-lane bonding (up to 8 lanes) for high bandwidth and scalability
- DRP support enabling real-time transceiver optimization and link diagnostics
- Deterministic low latency (lightweight Aurora protocol)
- Built-in data CRC support
- Automatic link initialization and maintenance
- Link health and error status monitoring
|
| Cooling |
|