Industrial GPU System | 120GQ
VINCANWO GROUP
| Availability: | |
|---|---|
| Quantity: | |
| Model | 120GQ |
| Machine Frame | 1U rack |
| CPU | Dual Socket P+ (LGA-4189) 3rd Gen Intel® Xeon |
| Chipset | Intel® C621A |
| RAM | 16pcs of memory slots max supports 6TB RECC 3200/2933 MHZ |
| Hard Disk Slot | 2pcs of NVME hot swappable hard drive slots |
| Hard Disk Specifications | 2.5" NVME hard disk |
| CPU card | Max supports 4pcs of GPU cards |
| Card RTX A6000, A40, A16, A100, RTX A5000, A30, A2 | |
| Network Card | 2pcs of RJ45 10GB |
| RAID | RAID 0,1,5,10 |
| Display Controller | 1 x VGA |
| External Device Ports | USB 3.0 |
| CD Driver | N/A |
| Cooling Action | 9pcs of high speed fans |
| Power Supply | 2000W high speed power supply 80PLUS certification |
| Management | IPMI 2.0 interface |
| Operating System | Windows, Linux |
| Temperature | Working Environment: - 10℃ ~ + 35℃ Non-working Environment: - 40℃ ~ + 70℃ |
| Humidity | Working Environment: 8%~90% Non-working Environment: 5%~95% |
| Product Size (mm) | 894 x 437 x 43 |
| Product Weight (kg) | 21.8 |
| Accessories | Installation guide rail / screw pack |
| Feature | 120GQ Specifications |
|---|---|
| GPU Compute | 8× NVIDIA RTX 6000 Ada(18,176 CUDA cores, 568 Tensor cores) |
| CPU | Dual Intel Xeon Scalable 4th Gen (56C/112T) |
| Memory | 1TB DDR5 ECC (4800MT/s) + 8× 48GB GDDR6 ECC (per GPU) |
| Acceleration | 3.4 PetaFLOPS FP8 (sparse), 1,360 TOPS INT8 (DL) |
| Rugged Rating | MIL-STD-901G (ship shock), IP67 (sealed chassis), -40°C to 55°C |
| Power | 2400W Titanium PSU (48V DC input w/ MIL-STD-1275/704) |
| Expansion | 8× Gen4 x16 PCIe(reinforced), 4× OCuLink-2(for co-processors) |
Extreme Environment Optimization
Vibration Control: Lord Corp. VS-100isolators (25Grms dampening)
Thermal Solution: 2-phase immersion cooling (3M Novec 7100) + external dry cooler
EMI Hardening: Schaffner FN2090 filters (CISPR 32 Class A)
GPU Performance Enhancements
Overclocking: Industrial-grade vapor chambers (sustained 2.5GHz boost)
Redundancy: GPU failover via NVIDIA Magnum IO
Safety Lock: Palladium-plated PCIe retainers (MIL-C-55302)
Deployment Flexibility
Modular Design: Slide-out GPU trays (tool-less swap <2min)
Mounting: 19" rack, DIN rail, or vehicle-mount (EN 50155)
| Interface | Application |
|---|---|
| Sensor Inputs | 8× CoaXPress-2 (12.5Gbps) for 10k cameras |
| Time-Sensitive Net | 4× TSN GbE (IEEE 802.1AS-2020, ≤250ns jitter) |
| Control Plane | Dual 100GbE (RoCEv2), CAN FD (8x ports) |
| Legacy Support | ARINC 429 (via M.2), 1553B (dual redundant) |
Autonomous Vehicles: 16-camera 3D perception @ 60fps (NVIDIA DRIVE OS)
Defense EW: 128-channel SIGINT processing (CUDA-accelerated GNU Radio)
Factory AI: 100-stream AOI with NVIDIA Metropolis
Simulation: Real-time CFD (Ansys Fluent GPU solver)
| Subsystem | Design |
|---|---|
| Immersion Cooling | Sealed tub w/ 3M Novec 7100, 35kW heat rejection |
| Power Conditioning | Vicor BCM6135 (48V→12V, 99% efficiency) |
| Backup Power | Maxwell K2 supercaps (5sec @ 2400W) |
| EMC Protection | TDK ZJYS51 ferrites on all I/O lines |
Vibration Mitigation
Naval Installations: Use Vibrasystems VLI-73 mounts per MIL-DTL-901E
Ground Vehicles: Enidine Hydraulic Shocks(EN 61373 Cat 1)
Thermal Management
Dry Cooler: Laird LC-480 (55°C ambient capability)
Coolant Maintenance: Annual Novec purity test (GC-MS analysis)
EMI/EMC Compliance
Cable Shielding: Gore GORE-FLIGHTcables w/ 360° MCRF connectors
Grounding: 10AWG braided straps to chassis (MIL-STD-188-124)
Bottom Line: The 120GQ delivers datacenter-grade GPU compute in a MIL-SPEC package, with immersion cooling for sustained performance in mobility/industrial environments. For deployment:
Validate MIL-STD-167-1 vibration profiles for naval applications
Implement deterministic Kubernetes (OpenShift RT) for control loops
Use NVIDIA DOCA for DPU-accelerated networking
| Model | 120GQ |
| Machine Frame | 1U rack |
| CPU | Dual Socket P+ (LGA-4189) 3rd Gen Intel® Xeon |
| Chipset | Intel® C621A |
| RAM | 16pcs of memory slots max supports 6TB RECC 3200/2933 MHZ |
| Hard Disk Slot | 2pcs of NVME hot swappable hard drive slots |
| Hard Disk Specifications | 2.5" NVME hard disk |
| CPU card | Max supports 4pcs of GPU cards |
| Card RTX A6000, A40, A16, A100, RTX A5000, A30, A2 | |
| Network Card | 2pcs of RJ45 10GB |
| RAID | RAID 0,1,5,10 |
| Display Controller | 1 x VGA |
| External Device Ports | USB 3.0 |
| CD Driver | N/A |
| Cooling Action | 9pcs of high speed fans |
| Power Supply | 2000W high speed power supply 80PLUS certification |
| Management | IPMI 2.0 interface |
| Operating System | Windows, Linux |
| Temperature | Working Environment: - 10℃ ~ + 35℃ Non-working Environment: - 40℃ ~ + 70℃ |
| Humidity | Working Environment: 8%~90% Non-working Environment: 5%~95% |
| Product Size (mm) | 894 x 437 x 43 |
| Product Weight (kg) | 21.8 |
| Accessories | Installation guide rail / screw pack |
| Feature | 120GQ Specifications |
|---|---|
| GPU Compute | 8× NVIDIA RTX 6000 Ada(18,176 CUDA cores, 568 Tensor cores) |
| CPU | Dual Intel Xeon Scalable 4th Gen (56C/112T) |
| Memory | 1TB DDR5 ECC (4800MT/s) + 8× 48GB GDDR6 ECC (per GPU) |
| Acceleration | 3.4 PetaFLOPS FP8 (sparse), 1,360 TOPS INT8 (DL) |
| Rugged Rating | MIL-STD-901G (ship shock), IP67 (sealed chassis), -40°C to 55°C |
| Power | 2400W Titanium PSU (48V DC input w/ MIL-STD-1275/704) |
| Expansion | 8× Gen4 x16 PCIe(reinforced), 4× OCuLink-2(for co-processors) |
Extreme Environment Optimization
Vibration Control: Lord Corp. VS-100isolators (25Grms dampening)
Thermal Solution: 2-phase immersion cooling (3M Novec 7100) + external dry cooler
EMI Hardening: Schaffner FN2090 filters (CISPR 32 Class A)
GPU Performance Enhancements
Overclocking: Industrial-grade vapor chambers (sustained 2.5GHz boost)
Redundancy: GPU failover via NVIDIA Magnum IO
Safety Lock: Palladium-plated PCIe retainers (MIL-C-55302)
Deployment Flexibility
Modular Design: Slide-out GPU trays (tool-less swap <2min)
Mounting: 19" rack, DIN rail, or vehicle-mount (EN 50155)
| Interface | Application |
|---|---|
| Sensor Inputs | 8× CoaXPress-2 (12.5Gbps) for 10k cameras |
| Time-Sensitive Net | 4× TSN GbE (IEEE 802.1AS-2020, ≤250ns jitter) |
| Control Plane | Dual 100GbE (RoCEv2), CAN FD (8x ports) |
| Legacy Support | ARINC 429 (via M.2), 1553B (dual redundant) |
Autonomous Vehicles: 16-camera 3D perception @ 60fps (NVIDIA DRIVE OS)
Defense EW: 128-channel SIGINT processing (CUDA-accelerated GNU Radio)
Factory AI: 100-stream AOI with NVIDIA Metropolis
Simulation: Real-time CFD (Ansys Fluent GPU solver)
| Subsystem | Design |
|---|---|
| Immersion Cooling | Sealed tub w/ 3M Novec 7100, 35kW heat rejection |
| Power Conditioning | Vicor BCM6135 (48V→12V, 99% efficiency) |
| Backup Power | Maxwell K2 supercaps (5sec @ 2400W) |
| EMC Protection | TDK ZJYS51 ferrites on all I/O lines |
Vibration Mitigation
Naval Installations: Use Vibrasystems VLI-73 mounts per MIL-DTL-901E
Ground Vehicles: Enidine Hydraulic Shocks(EN 61373 Cat 1)
Thermal Management
Dry Cooler: Laird LC-480 (55°C ambient capability)
Coolant Maintenance: Annual Novec purity test (GC-MS analysis)
EMI/EMC Compliance
Cable Shielding: Gore GORE-FLIGHTcables w/ 360° MCRF connectors
Grounding: 10AWG braided straps to chassis (MIL-STD-188-124)
Bottom Line: The 120GQ delivers datacenter-grade GPU compute in a MIL-SPEC package, with immersion cooling for sustained performance in mobility/industrial environments. For deployment:
Validate MIL-STD-167-1 vibration profiles for naval applications
Implement deterministic Kubernetes (OpenShift RT) for control loops
Use NVIDIA DOCA for DPU-accelerated networking