Unsolved
1 Rookie
•
1 Message
0
1299
December 18th, 2022 17:00
Can I install GPU Tesla T4 on PowerEdge R730
Hi all,
I have a PowerEdge R730 server. Now, we need to install the GPU to run HPC task. But currently, all the GPU in supported list are discontinued now, so I plan to use Tesla T4 for this task.
I had compare the Tesla T4 specs with one of the supported GPU (K40M), here is result
1. Dimensions
- K40M: 267mm, 2 slot vs T4: 168 mm, 1 slot
2. Interface
- All of them use PCIe 3.0x16
3. Power consumsion
- K40M: 245w vs T4: 70w
Are there others factor that may effect the installation? Can I install Tesla T4 on R730?
No Events found!


Dell-Dheeraj
143 Posts
0
December 19th, 2022 03:00
Hi @hungnvk54,
Tesla T4 is not a supported GPU configuration with the Dell PE R730. It's not that it won't work, but the said configuration is not tested and validated with the R730 server.
You can wait for a reply from the Dell community members who may have installed the Tesla T4 on R730.
Please refer to the below links which might help you.
https://dell.to/3BP82V9
https://dell.to/3FHHR45
Also, below are the list of GPU's validated on the r730.
fernanaguero
1 Rookie
•
7 Posts
0
August 16th, 2024 17:05
Hi @hungnvk54 any news on your experiments with the T4 on the R730? I am also considering adding one of these. Any suggestions or alerts? Thanks in advance!
coolnodje
1 Rookie
•
1 Message
•
2 Points
0
September 10th, 2025 11:44
For anyone that might be interested, I got it successfully working in a R640.
The heat generated by the GPU is taken into account and the fans speed up consistently to cool down the GPU when in use.
I'm wondering how it can work? Is it a server internal temp sensor doing the job? Or is the GPU sending information that can be taken into account by the server somehow ??
I have no doubt that plug in a Tesla T4 on a PCIe port in an R730 will work, but I'd like a confirmation that heat will be taken into account.
Dell-Martin S
Moderator
•
3.7K Posts
•
25.2K Points
1
September 10th, 2025 12:20
Hi,
How the GPU Communicates with the Server
For "validated" or officially supported GPUs: These communicate their status (including temperature) directly to the server's management controller, allowing the thermal management system to ramp up the fans based on actual GPU temperature data.
For "unsupported" or third-party GPUs: The server may only rely on nearby PCIe slot and board sensors to estimate GPU zone temperature, which is less precise. Sometimes, custom scripts or external tools (like fancontrol or ipmitool) are used to monitor actual GPU temps through NVIDIA tools and set fan speeds accordingly.
Maybe this thread is interesting for you
https://www.dell.com/community/en/conversations/poweredge-hardware-general/can-i-install-gpu-tesla-t4-on-poweredge-r730/647fa157f4ccf8a8de68cda9
(edited)
fernanaguero
1 Rookie
•
7 Posts
1
September 10th, 2025 14:26
Hi! Just chiming in to chase my previous post to confirm we have successfully installed an NVIDIA L4 GPU (passive cooling, half-height (low profile), half-length, single-slot) in the single x16 PCIe riser slot of the Dell R730 (SLOT4_G3_X16(CPU2). And everything is working nicely.