Proc10A 40/60 GigE Frame Grabber / Smart NIC

The Gidel Proc10A-40/60GigE smart NIC grabber family is ideal for high bandwidth applications, including multi-cameras and ultra-high bandwidth camera applications. Its high acquisition and processing capabilities ensures zero frame loss while offloading the CPU for key processing tasks. The board can support up to 6 x 10 GigE Vision cameras that connect directly to the board without a need for a switch. Using a switch, it is possible to grab from tens of cameras, e.g., 60 x 1 GigE. The board’s innovative InfiniVision architecture is optimized for multi-camera acquisition. All cameras’ data transferred synchronously to single buffer ready for immediate post processing.

The Proc10A-GigE family offers a number of options to accommodate diverse application needs, from a plug-and-play high-performance frame grabber to a fully customized system solution comprising acquisition, FPGA image processing, real-time image compression, multi-camera synchronization and complex I/O system control.

The Proc10A-GigE has an acquisition bandwidth of up to 60 Gb/s via a QSFP+ and two SFP+ connectors.

Target applications:

Machine Vision and Imaging
Volumetric video capture
Smart agriculture systems
Surveillance
High Performance Acquisition Systems

System specifications:

Camera Interface

4 or 6 x 10 GigE Links

Advanced Features

Link Aggregation mode Chunk Data Selective ROI acquisition Option for on-FPGA compression

Image Formats

Mono, Bayer, RGBA (8, 10, 12, 14 and 16 bits/color) and RGB (8, 10 and 12 bits/ color)

Max. Resolution

Horizontal: 16 K pixels (64-bit) Vertical: 65 K lines

Acquisition Rate

Up to 20 Gb/s

Host Bus

PCIe x8 Gen. 3

Frame Buffer

2 - 32 GB

Camera Types

Area and Line Scan

Form Factor

PCIe half length

Connectors

QSFP+ and 2 x SFP+

GPIO

12x LVTTL

OS Support

Win 10 and 11 Linux (kernel 2.6.x- 5.19) Linux version doesn’t include the ProcFG/InfiniVision GUI

The Proc10A-40/60 GigE Vision Grabber/Smart NIC

Contact us

Options:

Grabber architecture type

The Gidel frame grabber boards can operate in one of two unique frame grabber architectures, ProcFG or InfiniVision, by reprogramming the firmware. Grabber option #1: InfiniVision Enables using many synchronized cameras and/or to grab compressed images/videos. All the cameras' data, including from multiple boards, is organized in a single buffer. Each camera can have different resolution, and pixel format, and may be changed on the fly. The frame grabber can have one or more instances of InfiniVision, and all cameras of the same InfiniVision instance must be synchronized. Grabber option #2: ProcFG Each camera has a dedicated ProcFG instance. Frame size and pixel format must be fixed. Optimized for line scan cameras Option for grabbing ROIs. ProcFG may not be used for compressed data.

FPGA options

Optional on-FPGA processing using Gidel's libraries (IPs), Gidel's tailored library functions (IPs), or user customized IPs.

FPGA Type

Arria 10 270 | Arria 10 660 | Arria 10 1150

Proc10A-40/60 GigE are available in various variants with different FPGA. Contact us if you need assistance selecting the right model for your needs!

Proc10A FPGA resource comparison

Base model	Proc10A-40GigE	Proc10A-60GigE
FPGA	Intel Arria 10 GX 270 \| 660 \| 1150	Intel Arria 10 GX 660 \| 1150
FPGA-ALM	Up to 427,200	Up to 427,200
FPGA-M20K	Up to 2,713	Up to 2,713
FPGA - 18 x 19 Multiplier	Up to 3,036	Up to 3,036
Peripheral Memory:
DRAM A capacity	1 GB	1 GB
DRAM A sustain bandwidth	5.6 GB/s	5.6 GB/s
SoDIMM B capacity	16 GB	16 GB
SoDIMM B sustain bandwidth	8.96 GB/s	8.96 GB/s
SoDIMM C capacity	16 GB	16 GB
SoDIMM C sustain bandwidth	8.96 GB/s	8.96 GB/s

SDK:

The ProcFG and InfiniVision are each supported by its unique GUI and API suite for developing the user application. For description of ProcFG and InfiniVision, refer to the Options tab.

FPGA Programming Suite:

The Gidel grabber boards are supported by the ProcVision Suite enabling tailoring, debugging, and verification of the FPGA image processing and data flow.

Customization can be based on the InfiniVision or ProcFG flow, and can include multi-instances for supporting unique multi-camera topologies, as well as, image processing and compression.