Key Specifications
We couldn't find any available P620 GPUs. Search for alternative GPUs.
Frequently Asked Questions
Why choose the P620?
2GB GDDR5. Entry-level Pascal professional GPU for basic display.
When is the P620 not a good fit?
Only 2GB VRAM. A display adapter rather than a compute card.
Are P620 prices going up or down?
On-demand pricing has remained stable, averaging around $0.06/hr per GPU.
What size AI models can the P620 run?
With 2GB of VRAM, the P620 is mainly limited to small quantized models and lightweight GPU workloads.
How much VRAM does the P620 have?
The P620 has 2GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
What is the P620's memory bandwidth?
The P620 has 80 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
What data types does the P620 support?
The P620 supports 2 precision formats. Training: FP16, FP32.
Does the P620 support NVLink?
No. The P620 is a PCIe-only GPU with no NVLink, so it is better suited to single-GPU inference and smaller-scale workloads than large distributed training jobs.
Technical Specifications
| Feature | Specification |
|---|---|
| GPU Microarchitecture | Pascal |
| CUDA Cores | 512 |
| GPU Clock speed | 1266 MHz |
| Memory | 2GB GDDR5 |
| Memory Clock Speed | 1252 MHz |
| Memory Bus Width | 128 bit |
| Memory Bandwidth | 80 GB/s |
| FP32 (float) performance | 1.4 TFLOPS |
| FP16 (half) performance | 21.66 GFLOPS |
| FP64 (double) performance | 43.33 GFLOPS |
| TDP | 40W |
| System Interface | PCI Express Gen 3 x 16 |
| DirectX Support | 12 (12_1) |
| OpenGL Support | 4.6 |
| OpenCL Support | 3.0 |
| Vulkan Support | 1.3 |
| CUDA Compute Capability | 6.1 |
| Shader Model | 6.7 |
Source: official Nvidia P620 datasheet.
Last updated