Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Unable to set block size/thread via command args #11

Closed
calvintam236 opened this issue Sep 3, 2017 · 7 comments
Closed

Unable to set block size/thread via command args #11

calvintam236 opened this issue Sep 3, 2017 · 7 comments
Assignees
Milestone

Comments

@calvintam236
Copy link

calvintam236 commented Sep 3, 2017

I have run xmr-stak-nvidia, and it gives me around 100H/s using this config: "threads" : 46, "blocks" : 4, "bfactor" : 12, "bsleep" : 0.
Just build the xmrig-nvidia on macOS, and with auto gpu config on GeForce GT 750M @ 925/2508 MHz, I could only get around 60H/s with it gets around 16 threads..
I couldn't find an option to set block size/thread via command args. Anything you can help improve this result?

@calvintam236 calvintam236 changed the title Poor performance on GeForce GT 750M Unable to set block size/thread via command args Sep 3, 2017
@xmrig
Copy link
Owner

xmrig commented Sep 6, 2017

GPU options can changed only config file, place config.json near xmrig executable, it will update config and you can change options.
Thank you.

@xmrig xmrig added the question label Sep 6, 2017
@calvintam236
Copy link
Author

Can you add such command args if I use just one GPU to mine? (I don't want to setup json file because it is more complicated to edit it in docker...)

@xmrig xmrig mentioned this issue Sep 11, 2017
22 tasks
@xmrig
Copy link
Owner

xmrig commented Sep 19, 2017

@calvintam236 Can you suggest how command line options should named and how it use with multiple GPUs. Thank you.

@calvintam236
Copy link
Author

calvintam236 commented Sep 19, 2017

@xmrig As the miner is going to merge CPU, AMD, and Nvidia versions in the future, I would suggest to add prefix to command args now. For Nvidia version, you can use --cuda.
For multiple GPUs setup, how about --cuda-threads 1000,2000 --cuda-blocks 1000,2000, which 1000 for gpu index 0, 2000 for gpu index 1. Same changes needed for --bfactor and --bsleep.
--max-gpu-threads will be replaced by the new thread option?

@xmrig xmrig added this to the v2.4 milestone Oct 2, 2017
@xmrig xmrig self-assigned this Oct 2, 2017
xmrig added a commit that referenced this issue Oct 16, 2017
@xmrig
Copy link
Owner

xmrig commented Oct 16, 2017

I added new options:

  • --cuda-devices List of CUDA devices to use, optional, if not specified all available devices will be used. Examples: --cuda-devices=0, --cuda-devices=0,2, etc.
  • --cuda-launch=THREADSxBLOCKS Single option to specify threads and blocks, also can be comma separated. Other miners like ccminer and xmrMiner uses a launch option too.
  • --bfactor, --bsleep not renamed, no equivalent for AMD, can be comma separated now.
  • --max-gpu-threads deprecated, added --cuda-max-threads with equal functionality.

@calvintam236
Copy link
Author

@xmrig Thank you for the update and great work.

While --bfactor and --bsleep don't have equivalent for AMD, no cuda prefix will make user think that these options are available for AMD as well.

@calvintam236
Copy link
Author

Great job 👍 Now it's better

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

No branches or pull requests

2 participants