

🚀 Power your AI edge with Tesla P4 — where speed meets efficiency!
The NVIDIA Tesla P4 8GB GDDR5 Inferencing Accelerator is a passive-cooled, low-profile PCIe GPU designed for hyperscale AI workloads. Featuring 22 TOPS INT8 performance and 192 GB/s memory bandwidth, it dramatically reduces inference latency by 15X while delivering 60X better energy efficiency than traditional CPUs. Ideal for deep learning inference in scale-out servers, it balances high throughput with low power consumption.
| ASIN | B073V8MVK9 |
| Best Sellers Rank | #829 in Computer Graphics Cards |
| Customer Reviews | 2.6 2.6 out of 5 stars (6) |
| Date First Available | July 7, 2017 |
| Is Discontinued By Manufacturer | No |
| Item Weight | 15.2 ounces |
| Item model number | 900-2G414-0000-000 |
| Manufacturer | Nvidia Tesla |
| Package Dimensions | 10.1 x 7.1 x 3.2 inches |
C**Z
Did i get a bumb unit?
Not sure i'd dissuade potential buyers but maybe i got a bumb unit... This thing constantly over heats. I have run it in two different Dell R630's and unless the funs are over 60% the card thermal throttles. Even tried in in a T630 and results were worse (larger case more volume of airflow but less air pressure). While i was able to stabilize temps with R630 fan at 60-70% i feel that is quite an excessive requirement, not to mention excessively loud. Intake air temps were between 70-75F which is fine for ever server i have ever owned, maybe i need to run this card in the freezer.. lol.. Naw.. its just a bad unit. Sending it back and will look for an alternate solution.
S**S
VMware : Remote Desktop Works with NVidia!
We used the NVidia GPU for Windows Remote Desktop under VMware.
E**R
Tesla P4's are OK but outdated - buy T4's instead
Buy T4's instead for a similar price. The T4's are a newer, much superior product with double the memory. T4's are several times faster and include additional machine learning and ray tracing features.
Trustpilot
4 days ago
2 months ago