Nvidia P40
Budget legacy GPU when VRAM matters more than speed.
- Arch
- Pascal
- Memory
- 24 GB GDDR5
- Bandwidth
- 346 GB/s
- Released
- Q3 2016
Price history
Weekly median price per GPU per hour · Get the data
Aggregating historical prices...
At a glance
We don't currently track any P40 offers. See alternative GPUs.
Technical Specifications
Nvidia P40 · Per GPU
| Compute · dense | |
|---|---|
| INT8 | 47 TOPS |
| FP32 | 12 TFLOPS |
| Precision support | FP16FP32INT8 |
| Memory | |
|---|---|
| Capacity | 24 GB GDDR5 |
| Bandwidth | 346 GB/s |
| Bus width | 384-bit |
| ECC | Yes |
| Silicon | |
|---|---|
| Architecture | Pascal |
| Process | TSMC 16nm FinFET |
| Transistors | 11.8 billion |
| Shader cores | 3,840 CUDA cores |
| Compute units | 30 SMs |
| Fabric and host | |
|---|---|
| Host interface | PCIe 3.0 x16 |
| Power | |
|---|---|
| Board power | 250 W |
| Cooling | Passive |
Source: official Nvidia P40 datasheet.
Frequently Asked Questions
Why choose the P40?
24GB GDDR5 at very low cost. Legacy option for workloads that need more VRAM than the P4. Some availability from budget cloud providers.
When is the P40 not a good fit?
Pascal architecture with no mixed precision support. GDDR5 bandwidth is very low. Only viable for legacy batch inference where VRAM matters more than speed.
Alternatives to Nvidia P40
Last updated