A data center GPU against a workstation card on the Pascal architecture: 24 GB GDDR5 against 2 GB GDDR5 and 4.3x the memory bandwidth on the P40.
At a glance
Updated 34 minutes ago- Memory
- 24 GB
- GDDR5
- Memory bandwidth
- 346 GB/s
- From
- $0.09 /GPU/hr
- 1 provider
- Memory
- 2 GB
- GDDR5
- Memory bandwidth
- 80 GB/s
- From
- --
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | P40 | P620 | Difference |
|---|---|---|---|
| On-demand | $0.09 | -- | |
| Spot | $0.06 | No spot listings | |
| Median | $0.09 | -- |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
P40 vs P620 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | P40 | P620 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 24 GB GDDR5 | 2 GB GDDR5 | 12x more on P40 |
| Bandwidth | 346 GB/s | 80 GB/s | 4.3x higher on P40 |
| Bus width | 384-bit | 128-bit | 3x wider on P40 |
| Compute · dense | |||
| INT8 | 47 TOPS | Not supported | P40 only |
| FP32 | 12 TFLOPS | 1.3 TFLOPS | 9.2x higher on P40 |
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 250 W | 40 W | 6.2x higher on P40 |
| Released | Q3 2016 | Q1 2018 | 17 months newer on P620 |
| Precision | |||
| Supported formats | FP16FP32INT8 | FP16FP32 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: P40 datasheet and P620 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | P40 /GPU/hr | P620 /GPU/hr |
|---|---|---|
Vast.ai
|
On-Demand $0.09 | -- |
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia P40 was last updated on Sept. 7, 2026, and for Nvidia P620 on Sept. 7, 2026.