PNY NVIDIA A30X
SKU: NVBGPUNA30X24GB
NVIDIA converged accelerators are a unique and efficient architecture that combines NVIDIA GPUs, SmartNICs, and DPUs. They offer high performance and enhanced security for GPU accelerated workloads that need a lot of I/O, from the data centre to the edge.
The A30X is a converged accelerator that has an NVIDIA A30 Tensor Core GPU and a BlueField-2 DPU. It is ideal for use cases like 5G vRAN and AI-based cybersecurity that need a balance of compute and I/O performance. The A30X GPU can run multiple services with low latency and predictable performance, thanks to the onboard PCIe switch.
Available on back-order
Quantity:
Quantity:
System Configuration
Available on back-order
Details & Specifications
Details & Specifications
| Manufacturer | |
|---|---|
| Manufacturer Part Number |
NVA30XTCGPUCA-KIT |
Additional:
| Integrated DPU |
NVIDIA BlueField-2 |
|---|---|
| Media Engines |
1 Optical Flow Accelerator (OFA) |
| Multi-Instance GPU Support |
4 MIGs at 6 GB Each |
| Networking |
2x 100 Gbps ports, Ethernet or InfiniBand |
GPU:
| # Transistors |
54.2 Billion |
|---|---|
| FP32 Tensor Core (TFLOPS) |
82.6 | Sparsity |
| Peak FP16 Tensor Core Performance (TFLOPS) |
165 | Sparsity |
| Peak FP32 Performance (TFLOPS) |
10.3 |
| Peak FP64 Performance (TFLOPS) |
5.2 |
| Peak FP64 Tensor Core Performance (TFLOPS) |
10.3 | Sparsity |
| Peak INT8 Tensor Core Performance (TOPS) |
330 | Sparsity |
Memory:
| Memory Bandwidth (GB/s) |
1223 |
|---|---|
| Memory Capacity (GB) |
24 |
| NVLink |
Third-Generation | 200 GB/s Bidirectional |
| Memory Type |
HBM2e |
PCIe:
| PCIe Version |
PCIe 4.0 |
|---|
Power:
| # 8 Pin PCIe |
1 |
|---|---|
| Max Power Consumption (W) |
230 |
Software:
| API Support |
NVIDIA vCS (Virtual Compute Server) |
|---|---|
| Supported OS |
Linux, Windows |
Warranty:
| Mfr Warranty Length (M) |
36 |
|---|---|
| Mfr Warranty Type |
Direct To Manufacturer |
* Specifications for older discontinued variants can be view by clicking here