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
Runtime downloading of artifacts at module __init__ is a real pain for deployment, especially for parallel workloads. Would it be possible to move the lazy downloading of the artifact to Pkg.build()?
It would also be useful to have a mechanism to save these preferences so that they persist between sessions/versions (FFTW.jl and MPI.jl save their preferences to a file in .julia/prefs/, but JuliaLang/Pkg.jl#1835 will hopefully provide a more systematic way to do this).
Why not call CUDA.version() or something when you install the package? I'd rather not add even more functionality that should really live in Pkg here to support (pretty niche) use cases, the current set-up is complex enough already.
maleadt
changed the title
Resolve lazy artifact at Pkg.build time, save artifact preferences
More prominently document JULIA_CUDA_USE_BINARYBUILDER
Jun 8, 2020
Runtime downloading of artifacts at module
__init__
is a real pain for deployment, especially for parallel workloads. Would it be possible to move the lazy downloading of the artifact toPkg.build()
?It would also be useful to have a mechanism to save these preferences so that they persist between sessions/versions (FFTW.jl and MPI.jl save their preferences to a file in
.julia/prefs/
, but JuliaLang/Pkg.jl#1835 will hopefully provide a more systematic way to do this).cc: @vchuravy @jakebolewski
The text was updated successfully, but these errors were encountered: