github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-05 23:53:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2393] Inquiry on Optimal CPU and GPU Configurations for LLaMA 2(70B) #1392

New Issue
Closed
opened 2026-04-12 11:13:22 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 11:13:22 -05:00
Owner
Copy Link

Originally created by @gautam-fairpe on GitHub (Feb 7, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2393

Originally assigned to: @dhiltgen on GitHub.

I am currently exploring the capabilities of LLaMA 2 for various NLP tasks and am in the process of setting up the necessary hardware environment to ensure optimal performance. Given the complexity and resource-intensive nature of LLaMA 2(70B), I am seeking advice on the most suitable CPU and GPU configurations that can deliver the best performance for training and inference tasks with this model.

Specifically, I am interested in:

  • Recommendations for CPU and GPU models that are known to work well with LLaMA 2, considering both performance and cost-efficiency.
  • Any available estimation charts or benchmarks that illustrate the performance of LLaMA 2 with different hardware configurations. This information would be incredibly helpful for planning hardware investments and understanding the expected model throughput and latency.
Originally created by @gautam-fairpe on GitHub (Feb 7, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2393 Originally assigned to: @dhiltgen on GitHub. I am currently exploring the capabilities of LLaMA 2 for various NLP tasks and am in the process of setting up the necessary hardware environment to ensure optimal performance. Given the complexity and resource-intensive nature of LLaMA 2(70B), I am seeking advice on the most suitable CPU and GPU configurations that can deliver the best performance for training and inference tasks with this model. Specifically, I am interested in: - Recommendations for CPU and GPU models that are known to work well with LLaMA 2, considering both performance and cost-efficiency. - Any available estimation charts or benchmarks that illustrate the performance of LLaMA 2 with different hardware configurations. This information would be incredibly helpful for planning hardware investments and understanding the expected model throughput and latency.
GiteaMirror commented 2026-04-12 11:13:23 -05:00
Author
Owner
Copy Link

@igorschlum commented on GitHub (Feb 7, 2024):

Hi @gautam-fairpe I bought a Mac Station with Max memory. (192GB), si J cab Run simultaneously and locally all the LLM.

It's quick and I know that ai will always be able to sell it if I'm not more happy with it.

<!-- gh-comment-id:1933108515 --> @igorschlum commented on GitHub (Feb 7, 2024): Hi @gautam-fairpe I bought a Mac Station with Max memory. (192GB), si J cab Run simultaneously and locally all the LLM. It's quick and I know that ai will always be able to sell it if I'm not more happy with it.
GiteaMirror commented 2026-04-12 11:13:24 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (May 7, 2024):

For Llama2/3 70B you'd want at least 48GB of VRAM, if not closer to 64GB to fit the model and context size

<!-- gh-comment-id:2097128324 --> @jmorganca commented on GitHub (May 7, 2024): For Llama2/3 70B you'd want at least 48GB of VRAM, if not closer to 64GB to fit the model and context size
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#1392
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?