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

[core] Option to make chunks optional #778

Merged
merged 1 commit into from
Feb 5, 2020

Conversation

fabiencastan
Copy link
Member

When computing on a single machine with very large ressources, the chunks can limit the ressources used.
A new environment variable MESHROOM_USE_MULTI_CHUNKS can be set to 0 to disable it.

Warning: If you mix usage with and without this option, the nodes status will be incoherent.

@fabiencastan fabiencastan added this to the Meshroom 2020.1.0 milestone Feb 5, 2020
@natowi natowi mentioned this pull request Feb 5, 2020
@fabiencastan fabiencastan linked an issue Feb 5, 2020 that may be closed by this pull request
@fabiencastan fabiencastan merged commit 2e6990d into develop Feb 5, 2020
@fabiencastan fabiencastan deleted the dev/makeChunksOptional branch February 5, 2020 18:47
@natowi
Copy link
Member

natowi commented Feb 5, 2020

Thank you for this feature.

A new environment variable MESHROOM_USE_MULTI_CHUNKS can be set to 0 to disable it.

Like this?:


set MESHROOM_USE_MULTI_CHUNKS=0
set MESHROOM_INSTALL_DIR=%CD%
set PYTHONPATH=%CD% && python meshroom/ui

Will there be a checkbox in the Meshroom main menu?

@fabiencastan
Copy link
Member Author

Yes, that's it.

It could become part of the settings and integrated in the menus. But for now, it's really an advanced feature for specific needs and we need to measure the impact before going further.

@natowi
Copy link
Member

natowi commented Feb 5, 2020

we need to measure the impact before going further.

Reduces overall computation time by ~10% (tested with Monstree full), tested on i7-7820, GTX1070, 32GB
I will do a second test on a workstation.

@natowi
Copy link
Member

natowi commented Mar 12, 2020

I tested this on a 48 core workstation. The speed-up is not huge, it reduces the overall computation time by ~10%. This reduces computation time in FE, SfM, DM and M but increases computation time slightly in DMF and TEX.

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

Successfully merging this pull request may close these issues.

Bad multithreading?
2 participants