github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2882] Mixtral on Ollama, Nvidia RTX 3090 24G vs Nvidia A5000 24G : A Comparative Experience #27520

New Issue
Closed
opened 2026-04-22 04:55:48 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 04:55:48 -05:00
Owner
Copy Link

Originally created by @nejib1 on GitHub (Mar 2, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2882

Hello,

This is not an issue, it can be closed immediately, I just wanted to share my experience using Mixtral 26G on Ollama, comparing

Nvidia RTX 3090 and the Nvidia RTX A5000 on the same hardware, a SuperMicro 1028GR-TR server. Here's what I found:

Speed: I didn't notice any difference in speed, both GPUs perform similarly in this regard.

Temperature: The RTX 3090 runs significantly hotter compared to the A5000. Throughout my tests, the A5000 remained impressively cool.

Fan Speed: The RTX 3090's fans run at a minimum of 65%, whereas the A5000's fans operate at 30%, which is another point in favor of the A5000.

Power Consumption: The A5000 is more energy-efficient, consuming up to 230 watts at peak, whereas the RTX 3090 can consume up to 350 watts.

Conclusion: For AI purposes (Ollama in any case), The A5000 is clearly superior in all the aspects I tested.
It offers comparable speed with significantly better temperature management, lower fan speeds, and reduced power consumption.

Originally created by @nejib1 on GitHub (Mar 2, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2882 Hello, This is not an issue, it can be closed immediately, I just wanted to share my experience using Mixtral 26G on Ollama, comparing Nvidia RTX 3090 and the Nvidia RTX A5000 on the same hardware, a SuperMicro 1028GR-TR server. Here's what I found: **Speed:** I didn't notice any difference in speed, both GPUs perform similarly in this regard. **Temperature:** The RTX 3090 runs significantly hotter compared to the A5000. Throughout my tests, the A5000 remained impressively cool. **Fan Speed:** The RTX 3090's fans run at a minimum of 65%, whereas the A5000's fans operate at 30%, which is another point in favor of the A5000. **Power Consumption:** The A5000 is more energy-efficient, consuming up to 230 watts at peak, whereas the RTX 3090 can consume up to 350 watts. **Conclusion:** For AI purposes (Ollama in any case), **The A5000 is clearly superior** in all the aspects I tested. It offers comparable speed with significantly better temperature management, lower fan speeds, and reduced power consumption.
GiteaMirror commented 2026-04-22 04:55:49 -05:00
Author
Owner
Copy Link

@khu834 commented on GitHub (Mar 4, 2024):

How does Mixtral fit on either of those GPU?
Neither of them has 26G of GPU memory

<!-- gh-comment-id:1975460827 --> @khu834 commented on GitHub (Mar 4, 2024): How does Mixtral fit on either of those GPU? Neither of them has 26G of GPU memory
GiteaMirror commented 2026-04-22 04:55:50 -05:00
Author
Owner
Copy Link

@nejib1 commented on GitHub (Mar 4, 2024):

I have an A4000 already installed, but tests were by combining A4000+A5000 and A4000+RTX 3090

<!-- gh-comment-id:1975503844 --> @nejib1 commented on GitHub (Mar 4, 2024): I have an A4000 already installed, but tests were by combining A4000+A5000 and A4000+RTX 3090
GiteaMirror commented 2026-04-22 04:55:52 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (May 16, 2024):

We just got an A6000 Ada to test against our 4090s. Definitely agree it's a much better card, but it's also $3k more expensive :-D

<!-- gh-comment-id:2116364746 --> @pdevine commented on GitHub (May 16, 2024): We just got an A6000 Ada to test against our 4090s. Definitely agree it's a much better card, but it's also $3k more expensive :-D
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#27520
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?