Imagine a production line that freezes because the computer cannot process video or AI data fast enough. The Geniatech APC3588, 4GB/32GB, 4G LTE Octa-Core 8K AI Industrial Computer is built for those tough industrial environments. It gives you the power to run 8K video, AI inference, and multiple connections without lag. If you have been struggling with slow, unreliable controllers, this machine solves that problem and keeps your operation moving.
Geniatech APC3588, 4GB/32GB, 4G LTE Octa-Core 8K AI Industrial Computer
Many factories, smart cities, and transport hubs need a computer that can handle high-resolution video and AI models while staying online on a cellular network. When the device cannot keep up, data loss and downtime become costly. The APC3588 brings an octa-core RK3588 CPU, a 6TOPS NPU, and built-in LTE, so you get real-time processing and constant connectivity. It turns a fragile setup into a reliable edge node that never misses a frame or a command.
Key Features
- RockChip RK3588 octa-core 64-bit CPU up to 2.4 GHz.
- Mali-G610 MP4 GPU with OpenGL ES, Vulkan, and OpenCL support.
- 6 TOPS NPU compatible with TensorFlow, PyTorch, MXNet, Caffe.
- 4 GB LPDDR4/5 memory, expandable to 32 GB.
- 32 GB eMMC storage plus high-speed TF/SD slot up to 128 GB.
- Dual HDMI 2.1 outputs (8K @ 60 fps or 4K @ 120 fps) and one HDMI 2.1 input.
- Ethernet (1 GbE + 100 MbE), Wi-Fi 6, Bluetooth 5.0, optional LTE module.
- USB Type-C, USB 3.0 host, two USB 2.0 hosts, and OTG port.
- RS232, RS485, CAN, and GPIO interfaces for industrial control.
- Power via 12 V 2 A DC input or POE option up to 24 W.
Specifications
| Chip |
|
| Chip |
RockChip RK3588 |
| Market-oriented |
Global |
| Processors |
|
| CPU |
Octa-core 64-bit (4 × Cortex-A76 + 4 × Cortex-A55), 8nm advanced process up to 2.4GHz |
| GPU |
Mali-G610 MP4 quad-core GPU |
| GPU Features |
OpenGL ES 1.1/2.0/3.1/3.2, Vulkan 1.1, OpenCL 2.2, 450 GFLOPS |
| NPU |
NPU computing power up to 6TOPS, support INT4/INT8/INT16 hybrid computing |
| NPU Framework Support |
Can realize network model transformation based on
TensorFlow/MXNet/Pytorch/Caffe and other series of frameworks.
|
| ISP |
Integrated 48MP ISP and HDR & 3DNR multi-algorithm gas pedal |
| Codecs |
|
| Video decoding |
8K@60fps H.265/VP9/AVS2
8K@30fps H.264 AVC/MVC
4K@60fps AV1
1080P@60fps MPEG-2/-1/VC-1/VP8
|
| Video encoding |
8K@30fps encoding, support H.265 / H.264 |
| Multi-channel Support |
Up to 32-channel 1080P@30fps decoding and 16-channel 1080P@30fps encoding. |
| Storage |
|
| Memory |
4GB 64bit LPDDR4/LPDDR4x/LPDDR5 up to 32GB |
| Flash |
32GB eMMC |
| SD |
TF 8G-128G, 1-bit / 4-bit SD bus, support default, high speed,
UHS-I/SDR50, SDR104 transmission mode
|
| Network |
|
| Ethernet |
2 (1 × 1000 Mbps + 1 × 100 Mbps) |
| WiFi Module |
2.4G/5.8G WiFi6, BT5.0 |
| 4G Module |
LTE for EU, LTE for AU, LTE for US (Optional) |
| I/O Interface |
|
| HDMI® Output |
2 × HDMI®2.1 8K@60fps or 4K@120fps |
| HDMI® Input |
1 × HDMI®2.1 4K@60fps |
| USB |
1 × USB Type-C, 1 × USB Host3.0, 2 × USB Host2.0, 1 × USB OTG 2.0 |
| RS232 |
1 ±5~±15V |
| RS485 |
1 TIA485/EIA-485-A -7~+12V |
| CAN |
1 ISO11898-2:2016 and SAE J2284-1 to SAE J2284-5 ESD (8kV IEC and HBM) 5Mbit/s |
| GPIO |
4 3.3V level ±15KV (air) ±10KV (contact) |
| SIM |
1 Micro SIM Card |
| Operating System |
|
| OS |
Android12 (Optional Linux Ubuntu Desktop, Linux Ubuntu Server, Debian11.0, Buildroot, RTLinux) |
| Environmental |
|
| Temperature |
-20℃~60℃ |
| Mechanical Characteristics |
|
| Dimensions |
160mm × 125.5mm × 48mm |
| Housing |
Aluminum profile with fixing holes on the bottom sides |
| Power |
|
| DC IN |
DC Input Voltage 12V 2A (DC5.5 × 2.1mm) |
| POE |
POE Option Maximum power consumption: 24W |
