sales@north-bridges.com
86-13862120662
Note:
Product specifications on the website are for reference only, subject to the latest manufacturer-confirmed datasheet. Please contact us for official technical documents before project evaluation and quotation.
Operating system, BSP, drivers and SDK vary by hardware batch. Please contact us for official software compatibility documents before project development.
VPX GPU Accelerator Edge AI Computing Boards
This series of rugged VPX AI/GPU accelerator boards features diversified high-performance computing architectures in standard 3U/6U VPX form factors.
Wide-temperature design guarantees stable operation under extreme ambient conditions, supporting real-time video analysis, multi-sensor fusion and large-scale parallel data processing for aviation, industrial and mobile embedded applications.
Product Selction Table
| Part Number | Form Factor | Operating Temperature | Onboard Memory | AI Peak Computing Performance | Main Processing Chip | High-speed Backplane Interface | External Optical/IO Interface | Expansion Interface | Application Scenario |
| EL12PRCA5A001 | 6U VPX 5HP (Air Cooling) | 0℃ ~ +55℃ | FPGA: 8GB DDR4; GPU: Max 12GB LPDDR4 + Max 64GB eMMC | 20 TOPS @INT8 | XCVU3P + HUAWEI ATLAS- 200I-A2 | FPGA: PCIe3.0 x8 Serdes to P1; GPU: PCIe3.0 x4 + SPI to P1 | 1 x QSFP28, GigE, HDMI, USB3.0, MIPI DSI, SD Card, Type-C JTAG, SSMC | 1 × FMC+ HSPC (VITA57.4) | Edge AI inference, multi-sensor data fusion, remote sensing analysis, aviation intelligent monitoring, rugged parallel computing |
| EL12PRCA5A002 | 6U VPX 7HP (Air Cooling) | -43℃ ~ +65℃ | 4×24GB LPDDR4 + Max 64GB Emmc | 352 TOPS @INT8 (176 TFLOPS @FP16) | Dual HUAWEI ASCEND 310P | 2×PCIe4.0 x16 (Compatible with PCIe3.0/2.0 x8/x4) | 2 × Type-C Debug Ports, 1 Reset Switch, Indicators | 8 x HCCS Inter-chip | Real-time target recognition, radar signal processing, multi-channel video analysis, mobile embedded intelligent system |
| EL12PRCA5A003 | 6U VPX 5HP (Air Cooling) | 0℃ ~ +55℃ (Ventilated rack only) | 2×16GB GDDR6 (Dual RTX A4500), or 2×16GB GDDR5 (Dual Quadro P5000) | Single A4500: 19.35 TFLOPS @FP32); Single P5000: 6.49 TFLOPS @FP32 | NVIDIA dual RTX-A4500 or dual Quadro-P5000 | 2×PCIe3.0x8 (P1, P5), 1×PCIe3.0x16 (P2), 4×DP (P6) | Front Panel: 2×VGA, 2×DP | Dual MXM3.1 | Radar signal processing, 3D simulation, remote sensing image analysis, offline AI training, aviation & mobile rugged computing |
| EL12PRCA5B003 | 6U VPX 10HP (Conduction Cooling) | -40℃ ~ +55℃ (Continuous full-load operation supported in sealed rugged chassis) | |||||||
| EL12PRCA5A004 | 6U VPX 5HP (Conduction Cooling) | -20℃ ~ +60℃ | 16GB HBM2E (MR-V50), or 32GB HBM2E (MR-V100) | MR-V100: 295 TOPS @INT8; MR-V50: 192 TOPS @INT8 | China ILUVATAR MR-V100 or MR-V50 | PCIe4.0 x16/x8 (Backward Compatible with PCIe 3.0/2.0/1.0) | 2 × HDMI, 2 × DVI, 1×DP | Single MXM3.1 Type-A | Multi-modal data analysis, edge intelligent detection, real-time image processing, industrial rugged embedded computing |
| EL12PRCA5A005 | 6U VPX 5HP (Air Cooling) | 0℃ ~ +55℃ | 8GB GDDR6 (RTX3070), or 16GB GDDR5 (P5000M), or 8GB GDDR5 (GTX1070) | RTX3070: 20.3 TFLOPS @FP32; P5000M: 6.49 TFLOPS @FP32; GTX1070M: 7.42 TFLOPS @FP32 | NVIDIA RTX3070, or Quadro- P5000M or GTX1070M | PCIe 3.0 ×8 | 2 × HDMI, 2 × DVI, 1×DP | Single MXM3.1 | Machine vision, parallel signal computation, geographic mapping, industrial simulation, rugged embedded data processing |
| EL12PRCA5A006 | 3U VPX 5HP (Air Cooling) | -20℃ ~ +60℃ (Ventilated rack only, intermittent full-load） | 32GB LPDDR5 + 64GB eMMC or 64GB LPDDR5 + 64GB eMMC | 64GB/32GB Version: 275/200 TOPS @INT8 | NVIDIA Jetson AGX Orin 32GB/64GB | PCIe3.0 x8, PCIe2.0 x1; 4×GigE, 2×SATA, RS232, RS422, 4×GPIO | Front Panel: 1×HDMI, 1×USB2.0, 1×USB3.0, 1×GigE, J30J Debug Port | M.2 NVMe | Multi-channel video analysis, real-time target detection, edge AI inference, aviation intelligent perception, embedded vision |
| EL12PRCA5B006 | 3U VPX 5HP (Conduction Cooling) | -40℃ ~ +80℃ (Continuous full-load in sealed rugged chassis） | |||||||
| EL12PRCA5A007 | 3U VPX 5HP (Air Cooling) | 0℃ ~ +55℃ (Ventilated rack only, intermittent full-load operation) | 12GB GDDR6 | 10.94 TFLOPS @FP32 | NVIDIA RTX3060 | PCIe 3.0 x8 (P1); 4×DP Reserved (P6) | Front Panel: 3 × DP High- definition Display Ports | Single MXM3.1 Type-B | High-speed image rendering, machine vision inspection, industrial edge computing, rugged display & data processing |
| EL12PRCA5B007 | 3U VPX 10HP (Conduction Cooling) | 0℃ ~ +55℃ (Continuous full-load operation supported in sealed rugged chassis) | |||||||
| EL12PRCA5A008 | 3U VPX 4HP (Conduction Cooling) | -25℃ ~ +80℃ | 8GB LPDDR4 +16GB eMMC or 16GB LPDDR4 +16GB eMMC | 21 TOPS @INT8 | NVIDIA Jetson Xavier NX | 1×GigE, 1×RS422 | Front Panel: 1×HDMI, 2×Micro USB2.0, 1×Type-C Debug Port | M.2 NVMe | Lightweight edge AI detection, multi-sensor fusion, mobile terminal monitoring, embedded real-time vision analysis |
| EL12PRCA5A009 | 3U VPX 5HP (Air Cooling) | 0℃ ~ +55℃ | 32GB LPDDR4 + 32GB eMMC | 32 TOPS @INT8 | NVIDIA Jetson AGX Xavier | PCIe3.0x8; 4×GigE, 2×SATA, RS232, RS485, 4×GPIO, SPI, I2C | Front Panel: 1×HDMI, 2×USB2.0/3.0, 1×RS232 | M.2 NVMe | Real-time video analytics, target tracking, industrial IoT edge computing, aviation embedded intelligent system |
| EL12PRCA5A010 | 3U VPX 5HP (Conduction Cooling) | -40℃ ~ +70℃ | 4GB LPDDR4 + 16GB eMMC | 1 TFLOPS @FP32 | NVIDIA Jetson TX2 | PCIe3.0 Switch with PCIe2.0x4 +PCIe1.0, 2×GigE, 1×RS422 | 1×HDMI, 1×USB2.0 OTG, 2×USB3.0, 6×MIPI CSI-2 Camera Input | Compatible with RTG004 RTM Rear Transition Module | Low-power embedded data acquisition, small-scale image recognition, mobile terminal control, portable rugged computing |
PN.: EL12PRCA5A001
XCVU3P + HUAWEI ATLAS200IA2 6U VPX Rugged Heterogeneous
GPU Accelerator FPGA Signal Processing Platform
Specifications
| Form Factor | 6U VPX, 5HP | ||||||||
| Chip | XCVU3P + HUAWEI ATLAS-200I-A2 | ||||||||
| AI Peak Performance | 20 TOPS@INT8 | ||||||||
| Memory | FPGA: 8GB DDR4 | ||||||||
| GPU: Max 12GB LPDDR4 + Max 64GB eMMC | |||||||||
| High-speed Interface | FPGA: PCIe3.0 x8 Serdes to P1 | ||||||||
| GPU: PCIe3.0 x4 + SPI to P1 | |||||||||
| Optical/IO Interface | 1 x QSFP28, GigE, HDMI, USB3.0, MIPI DSI, SD Card, Type-C JTAG, SSMC | ||||||||
| Expansion Interface | 1 × HSPC, VITA57.4 FMC+ | ||||||||
| Operating Temp | 0℃ ~ +55℃ | ||||||||
Block Diagram
PN.: EL12PRCA5A002
Dual HUAWEI ASCEND-310P 6U VPX Heterogeneous Rugged
GPU Accelerator Edge AI Computing Platform
Specifications
| Form Factor | 6U VPX, 7HP | ||||||||
| Chip | Dual HUAWEI ASCEND 310P | ||||||||
| AI Peak Performance | 352 TOPS@INT8, 176 TOPS@FP16 | ||||||||
| Memory | 96GB LPDDR4 | ||||||||
| High-speed Interface | 2 × PCIe 4.0 x16 (Compatible with PCIe 3.0/2.0 x8/x4) | ||||||||
| Optical/IO Interface | 2 × Type-C Debug Ports, 1 Reset Switch, Indicators | ||||||||
| Expansion Interface | Inter-NPU: 8 x HCCS | ||||||||
| Operating Temp | -43℃ ~ +65℃ | ||||||||
Block Diagram
PN.: EL12PRCA5A003 / EL12PRCA5B003
Dual MXM3.1 6U OpenVPX Rugged
GPU Accelerator Edge AI Computing Platform
(Support NVIDIA RTX A4500 / Quadro P5000)
Specifications
| Form Factor | 6U OpenVPX, 5HP (Air Cooling) / 10HP (Conduction Cooling) | ||||||||
| Chip | Dual NVIDIA RTX-A4500 or Quadro-P5000 | ||||||||
| AI Peak Performance | Single A4500: 19.35 TFLOPS@FP32 | ||||||||
| Single P5000: 6.49 TFLOPS@FP32 | |||||||||
| Memory | Dual A4500: 2×16GB GDDR6 | ||||||||
| Dual P5000: 2×16GB GDDR5 | |||||||||
| High-speed Interface | 2×PCIe3.0 x8 (P1, P5), 1×PCIe3.0 x16 (P2), 4×DP (P6) | ||||||||
| Optical/IO Interface | Front Panel: 2×VGA, 2×DP | ||||||||
| Expansion Interface | Dual MXM3.1 | ||||||||
| Operating Temp | 5HP: 0℃ ~ +55℃ | ||||||||
| 10HP: -40℃ ~ +55℃ | |||||||||
Block Diagram
PN.: EL12PRCA5A004
Single MXM3.1 Type-A 6U VPX Rugged
GPU Accelerator AI Parallel Computing Platform
(Support China ILUVATAR MR-V100 / MR-V50)
Specifications
| Form Factor | 6U VPX, 5HP | ||||||||
| Chip | China ILUVATAR MR-V100 or MR-V50 | ||||||||
| AI Peak Performance | MR-V100: 295 TOPS@INT8 | ||||||||
| MR-V50: 192 TOPS@INT8 | |||||||||
| Memory | MR-V100: 32GB HBM2E | ||||||||
| MR-V50: 16GB HBM2E | |||||||||
| High-speed Interface | PCIe 4.0 x16/x8 (Backward Compatible with PCIe3.0/2.0/1.0) | ||||||||
| Optical/IO Interface | Front Panel: 1 × HDMI | ||||||||
| Expansion Interface | Single MXM3.1 Type-A Slot | ||||||||
| Operating Temp | -20℃ ~ +60℃ | ||||||||
Block Diagram
PN.: EL12PRCA5A005
Single MXM3.1 6U VPX Rugged
GPU Accelerator Edge AI Computing Platform
(Support NVIDIA RTX 3070 / Quadro P5000M / GTX1070M)
Specifications
| Form Factor | 6U VPX, 5HP | ||||||||
| Chip | NVIDIA RTX3070 or Quadro P5000M or GTX1070M | ||||||||
| AI Peak Performance | RTX3070: 20.3 TFLOPS@FP32 | ||||||||
| P5000M: 6.49 TFLOPS@FP32 | |||||||||
| GTX1070M: 7.42 TFLOPS@FP32 | |||||||||
| Memory | RTX3070: 8GB GDDR6 | ||||||||
| P5000M: 16GB GDDR5 | |||||||||
| GTX1070M: 8GB GDDR5 | |||||||||
| High-speed Interface | PCIe 3.0 ×8 | ||||||||
| Optical/IO Interface | Front Panel: 1 × HDMI | ||||||||
| Expansion Interface | Single MXM3.1 slot | ||||||||
| Operating Temp | 0℃ ~ +55℃ | ||||||||
Block Diagram
PN.: EL12PRCA5A006 / EL12PRCA5B006
NVIDIA Jetson-AGX-Orin 3U OpenVPX Rugged
GPU Accelerator Edge AI Computing Platform
Specifications
| Form Factor | 3U OpenVPX, 5HP (Air Cooling) / 5HP (Conduction Cooling) | ||||||||
| Chip | NVIDIA Jetson AGX Orin (Optional 64/32 GB Version) | ||||||||
| AI Peak Performance | 64 / 32 GB Version: 275 / 200 TOPS@INT8 | ||||||||
| Memory | 64 / 32 GB LPDDR5 | ||||||||
| High-speed Interface | PCIe3.0 x8, PCIe2.0 x1; 4×GigE, 2×SATA, RS232, RS422, 4×GPIO | ||||||||
| Optical/IO Interface | Front Panel: 1×HDMI, 1×USB2.0,1×USB3.0,1×GigE, J30J | ||||||||
| Expansion Interface | M.2 NVMe | ||||||||
| Operating Temp | 5HP (Air Cooling): -20℃ ~ +60℃ | ||||||||
| 5HP (Conduction Cooling): -40℃ ~ +80℃ | |||||||||
Block Diagram
PN.: EL12PRCA5A007 / EL12PRCA5B007
Single MXM3.1 Type-B 3U OpenVPX Rugged
GPU Accelerator Edge AI Computing Platform
(Support NVIDIA RTX 3060)
Specifications
| Form Factor | 3U OpenVPX, 5HP (Air Cooling)/10HP (Conduction Cooling) | ||||||||
| Chip | NVIDIA RTX 3060 | ||||||||
| AI Peak Performance | 10.94 TFLOPS@FP32 | ||||||||
| Memory | 12GB GDDR6 | ||||||||
| High-speed Interface | PCIe 3.0 x8 (P1); 4×DP Reserved (P6) | ||||||||
| Optical/IO Interface | Front Panel: 3 × DP Display Ports | ||||||||
| Expansion Interface | Single MXM3.1 Type B | ||||||||
| Operating Temp | 5HP: 0℃ ~ +55℃ (Ventilated rack only, intermittent full-load operation) | ||||||||
| 10HP: 0℃ ~ +55℃ (Continuous full-load operation) | |||||||||
Block Diagram
PN.: EL12PRCA5A008
NVIDIA Jetson-Xavier-NX 3U VPX Rugged Low-Power
GPU Accelerator Edge AI Computing Platform
Specifications
| Form Factor | 3U VPX, 4HP | ||||||||
| Chip | NVIDIA Jetson Xavier NX | ||||||||
| AI Peak Performance | 21 TOPS@INT8 | ||||||||
| Memory | 16GB or 8GB LPDDR4 | ||||||||
| High-speed Interface | 1×GigE, 1×RS422 | ||||||||
| Optical/IO Interface | Front Panel: 1×HDMI, 2×Micro USB2.0, 1×Type-C Debug Port | ||||||||
| Expansion Interface | M.2 NVMe | ||||||||
| Operating Temp | -25℃ ~ +80℃ | ||||||||
Block Diagram
PN.: EL12PRCA5A009
NVIDIA Jetson-AGX-Xavier 3U VPX Rugged
GPU Accelerator Edge AI Computing Platform
Specifications
| Form Factor | 3U VPX, 5HP (Air Cooling) | ||||||||
| Chip | NVIDIA Jetson AGX Xavier | ||||||||
| AI Peak Performance | 32 TOPS@INT8 | ||||||||
| Memory | 32GB LPDDR4x | ||||||||
Block Diagram
PN.: EL12PRCA5A010
NVIDIA Jetson-TX2 3U OpenVPX Rugged
GPU Accelerator Edge AI Computing Platform
Specifications
| Form Factor | 3U OpenVPX, 5HP (Conduction Cooling) | ||||||||
| Chip | NVIDIA Jetson TX2 | ||||||||
| AI Peak Performance | 1 TFLOPS | ||||||||
| Memory | 4GB LPDDR4 | ||||||||
Block Diagram
Ordering
| PN: | All Above | ||||||||
| Lead time: | Inquire for official quotation | ||||||||
| Custom designs: | Available upon request | ||||||||
| Support: | FPGA reference design, schematic review, driver development (quoted separately) | ||||||||
Contact: Windy Lian
Phone: 13862120662
E-mail: sales@north-bridges.com
Whatsapp:
Add: Suzhou City, China
We chat
