You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Although the circular dependency warning is not fatal its at least annoying and might have consequences, so I spent some time tracking it down (adding packages in Project.toml order manually). Here are the results from a non GPU Windows 10 processor:
Its source is the CUDA package and is the subject of CUDA.jl issue #2586. It is an "upstream bug" apparently fixed in Julia v1.11.2 and v1.10.8. This latter version is to be released in a week or so from 14th January 2025 and is currently not available.
I didn't need to go far with the search.
I ran v1.11.2 and added CUDA v5.6.0 and the warning ideed did not appear. The resulting system has Atomix 1.0.1 which appears in an error message which I will investigate once I am back at a GPU machine.
The text was updated successfully, but these errors were encountered:
As noted there, nobody except Julia developers can eliminate the warning, so there's no action to take. But we can monitor whether this is fixed by 1.10.8.
We're seeing weird stuff with 1.11, here is a PR about moving the tests to 1.11: CliMA/Oceananigans.jl#3836
Although the circular dependency warning is not fatal its at least annoying and might have consequences, so I spent some time tracking it down (adding packages in Project.toml order manually). Here are the results from a non GPU Windows 10 processor:
Its source is the CUDA package and is the subject of CUDA.jl issue #2586. It is an "upstream bug" apparently fixed in Julia v1.11.2 and v1.10.8. This latter version is to be released in a week or so from 14th January 2025 and is currently not available.
I didn't need to go far with the search.
I ran v1.11.2 and added CUDA v5.6.0 and the warning ideed did not appear. The resulting system has Atomix 1.0.1 which appears in an error message which I will investigate once I am back at a GPU machine.
The text was updated successfully, but these errors were encountered: