Skip to content
This repository has been archived by the owner on Apr 24, 2022. It is now read-only.

Work around for Maxwell Tesla m40 GPU slowdown? #2116

Open
Visual-Synthesizer opened this issue Jan 13, 2021 · 6 comments
Open

Work around for Maxwell Tesla m40 GPU slowdown? #2116

Visual-Synthesizer opened this issue Jan 13, 2021 · 6 comments

Comments

@Visual-Synthesizer
Copy link

Visual-Synthesizer commented Jan 13, 2021

Only getting 2.4Mh/s with each NVIDIA Tesla M40 GPU with 24GB RAM in my AI server. Not sure if this is bug, after much digging found other mentions of slowdowns for Compute 5.2 and maxwell GPU wich my GPU runs. Same gpu as quadro m6000.

Running Arch Linux, and latest version of Ethminer .19 via the AUR https://aur.archlinux.org/packages/ethminer-cuda/

Any tips greatly appreciated. Have exhausted all my google search and search on this forum. Thinking about trying older version?

Running:
ethminer -U -P ...........

she works but very slow. NVIDIA-SMI watch shows 100% GPU util but power usage only 90w out of 250w. Runs great on my other system with 1070 and 2080ti! thanks.

@Visual-Synthesizer
Copy link
Author

Visual-Synthesizer commented Jan 13, 2021

Hello, is there a charge for this subscription? Thank you. DaniJoy notifications@github.com 於 2021年1月13日 週三 11:41 寫道:

Only getting 2.4MS/s with each NVIDIA Tesla M40 GPU with 24GB RAM in my AI server. Not sure if this is bug, after much digging found other mentions of slowdowns for cc 5.2 and maxwell GPU wich my GPU runs. Running Arch Linux, and latest version of Ethminer .19 via the AUR https://aur.archlinux.org/packages/ethminer-cuda/ Any tips greatly appreciated. Have exhausted all my google search and search on this forum. Thinking about trying older version? Running: ethminer -U -P ........... she works but very slow. NVIDIA-SMI watch shows 100% GPU util but power usage only 90w out of 250w. Runs great on my other system with 1070 and 2080ti! thanks. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2116>, or unsubscribe

No, its free. you just have to buy all the parts, assemble them, program the Linux OS and pay for the power :-)

@KevinGoehler
Copy link

Did you find a solution?
I'm only getting 1.08Mh on my M6000 (12GB)
I'm using v0.18

@AGenchev
Copy link

AGenchev commented May 18, 2021

2.4MH are ~ 8USD per month of income.
How much did you pay for this 24GB M40 ?
I wonder whether to buy also M40 24G

@bobobobo9
Copy link

Did you find a solution?
I'm only getting 1.08Mh on my M6000 (12GB)
I'm using v0.18
hey bro u try some old driver like cuda 6.5 and old version ethminer?

@bobobobo9
Copy link

bobobobo9 commented May 19, 2021

Did you find a solution?
I'm only getting 1.08Mh on my M6000 (12GB)
I'm using v0.18
https://developer.nvidia.com/cuda-toolkit-archive find some old driver
https://www.nvidia.com/Download/Find.aspx or this pls tell me test reuslt
and maybe try command "-G" use opencl to mining

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

5 participants
@AGenchev @KevinGoehler @bobobobo9 @Visual-Synthesizer and others