Specifications
CPU
14th Gen Intel Core i9/i7/i5/i3 (up to 24-core), max 65 W TDP
Chipset
Intel R680E
AI Accelerator
Neuchips RecAccel N3000-series PCIe AI card
AI Performance
206 TOPS (INT8)
Accelerator Memory
LPDDR5 64 GB
LLM Support
Llama3-8B, Llama2-7B, Mistral-7B, TAIDE, Breeze-7B, Phi-2
RAG
Hardware-enabled with engine in chip
System Memory
2x DDR5-5200 SO-DIMM, up to 96 GB
Storage
2x front-access 2.5" SSD tray; M.2 Key M 2280 (PCIe Gen4); 2x SATA III RAID 0/1
Display
2x DisplayPort (up to 8K@60), 1x HDMI, 1x DVI-I
USB
6x USB 3.2 Gen2 Type-A; 1x USB 3.2 Gen2x2 Type-C (20 Gbps, 15 W)
Expansion
1x Mini PCIe; 2x M.2 B-Key; 1x M.2 E-Key
Serial / DIO
4x COM RS-232/422/485; 16-ch isolated GPIO
Networking
2x GigE (Intel I219LM with iAMT + Intel I210)
Power Input
DC 12-50 V, 16-mode software ignition control
Security
Infineon SLB9670 TPM 2.0; Intel vPro, TCC, TSN
Mechanical
260 x 240 x 79 mm; 4.91 kg
Operating Temperature
-20 C to +45 C (with system fan)
Certifications
CE, FCC, EN50155, EN50121-3-2; MIL-STD-810G shock and vibration
SKU
ECX-3100 RAG
Overview
The Vecow ECX-3100 RAG is a compact, workstation-grade AI inference system built for retrieval-augmented-generation (RAG) and large-language-model workloads at the edge. It blends a 14th Gen Intel Core i9/i7/i5/i3 platform (up to 24 cores, Intel R680E chipset) with a Neuchips RecAccel N3000 PCIe accelerator card rated at 206 TOPS INT8 with 64 GB of LPDDR5.
Unusually for an industrial computer, RAG is hardware-enabled with an engine in the accelerator chip, and the system is validated for models including Llama3-8B, Llama2-7B, Mistral-7B, TAIDE, Breeze-7B and Phi-2. Wide 12-50 V power, ignition control and MIL-STD-810G shock/vibration ratings make it as comfortable in a vehicle or cabinet as in a rack.
Key Benefits
206 TOPS INT8 in a compact chassis brings local LLM inference to the edge. Hardware-enabled RAG accelerates retrieval workflows natively. Workstation-grade 14th Gen Intel Core and up to 96 GB DDR5 handle pre/post-processing, and wide 12-50 V input with ignition control and MIL-STD-810G ruggedness suit vehicle and industrial deployments.
Applications
On-premises LLM and RAG inference, conversational AI appliances, machine vision and video analytics, intelligent transportation and in-vehicle AI, industrial automation, and edge data-processing in factory or field cabinets.
Request a Quote — VECOW ECX-3100 RAG — RAG-ENABLED LLM AI INFERENCE WORKSTATION SYSTEM
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →