Shipping IP
Production Proven

CoaXPress IP Core

Host & Device • 1× / 2× / 4× lanes • AXI Ready

A synthesizable, lane-scalable CXP2.0 implementation with clean AXI interfaces, deterministic timing, and a complete verification environment built on SystemVerilog Bus Functional Models (BFMs).

Turnkey reference designs included

ZCU106 · KCU116 · Aper-Oculus carrier — build scripts and constraints ready to run.

CXP 2.0
Protocol
1× / 2× / 4×
Lane Modes
Host & Device
Configurations
PHY × Lanes
Lane 1
Lane 2
Lane 3
Lane 4
Link Layer & Lane Bonding
Control (AXI4-Lite)
Registers
Events
Triggers
Video Path (AXI4-Stream)
DMA

FPGA‑Focused CoaXPress

Host and Device Configurations with lane‑bonded throughput up to 4× CXP‑12. Clean AXI interfaces, deterministic timing, and plug‑and‑play DMA integration.

CXP Host & Device

Single code base supports both Host (frame grabber) and Device (camera) Configurations with build-time switchable top-levels.

1× / 2× / 4× Lanes

Scales from single-lane CXP-12 to quad-lane aggregations. Includes lane bonding, alignment, and deterministic skew compensation.

AXI-Stream Ready

Image and control paths expose clean AXI4-Stream/AXI4-Lite interfaces for DMA engines like XDMA or LitePCIe.

Production Proven

Delivered and tested on Xilinx ZCU106, KCU116, and our Aper-Oculus carrier with working camera/host reference designs.

Why this IP

Throughput

Aggregate up to 50 Gbps with low latency and alignment across lanes.

Determinism

Cycle‑accurate state machines and synchronizers for reproducible behavior.

Modularity

Composable subcores: PHY shim, link, control, and video packing layers.

Ready to integrate CoaXPress?

Ship faster with a verified, lane‑scalable IP core and working designs for ZCU106, KCU116, and Aper‑Oculus.

BFM testbenches included
Host & Device builds
1× / 2× / 4× lanes