Use the same default CPU target as PackageCompiler#47
Open
visr wants to merge 1 commit into
Open
Conversation
Member
That problem is now fixed on May be worth testing again to see if your issue is gone |
visr
force-pushed
the
default_cpu_target
branch
from
November 29, 2025 18:33
c124d97 to
ece781c
Compare
Author
|
I can't really tell, on Julia nightly both main and this PR fail in the same way. Full test output
|
Author
|
I tried 1.13.0-alpha2 and got the same error as on nightly above. However 1.12-nightly, (the future 1.12.3), which includes your fix,, passes all tests on this branch for me, so that seems promising. |
This was referenced Dec 3, 2025
visr
force-pushed
the
default_cpu_target
branch
from
December 29, 2025 10:14
ece781c to
067bae5
Compare
Author
|
Would be great if CI can be approved for this. I rebased, and now that Julia 1.12.3 is out I expect this to work (it does locally). Binaries get a bit bigger. Not sure how to best test this. |
visr
force-pushed
the
default_cpu_target
branch
from
January 19, 2026 14:47
067bae5 to
43ede0d
Compare
Author
|
This is working well for me, using it without trimming as a replacement for PackageCompiler. The advantage is that our build time went from 1h40m to 1h8m in Deltares/Ribasim#2761. |
visr
added a commit
to Deltares/Ribasim
that referenced
this pull request
Jan 27, 2026
Trying #2629 again now that we are some patch releases further along. EDIT: the switch to Julia v1.12 is now done with #2815, I rebased this to only add the JuliaC switch, which is less urgent. For JuliaC we use this branch: JuliaLang/JuliaC.jl#47. That should set the desired CPU_TARGET, which was the main reason for reverting the previous attempt. Status: I can run the TeamCity JuliaC generated binaries locally, so the CPU_TARGET seems to work. The Linux binaries seem to for some models. The libribasim.dll file grew from 0.9 to 1.2 GB. Update: this now works well on Windows. There is a libribasim.dll.a added to the zip root that we perhaps should remove, but the Windows build went from 1h40m to 1h8m, and testing the binaries from 20m to 17m as well so latency doesn't seem increased. I ran the Windows binaries locally. The binary wasn't working directly for me on Linux only after fiddling with the load path, we need to fix that, perhaps in main.rs or with a JuliaC rpath option or such. EDIT: see JuliaLang/JuliaC.jl#109 <details><summary>Details, now fixed with rpath=`@bundle`</summary> <p> ``` visser_mn@v-h7rdp002 ~/bin $ ribasim/bin/ribasim ~/ribasim-nl/trivial/ribasim.toml Failed to load libribasim from "/u/visser_mn/bin/ribasim/bin/../lib/libribasim.so" Error: DlOpen { desc: "libjulia.so.1.12: cannot open shared object file: No such file or directory" } visser_mn@v-h7rdp002 ~/bin $ export LD_LIBRARY_PATH=~/bin/ribasim/lib:$LD_LIBRARY_PATH visser_mn@v-h7rdp002 ~/bin $ ribasim/bin/ribasim ~/ribasim-nl/trivial/ribasim.toml Failed to load libribasim from "/u/visser_mn/bin/ribasim/bin/../lib/libribasim.so" Error: DlOpen { desc: "libjulia-internal.so.1.12: cannot open shared object file: No such file or directory" } visser_mn@v-h7rdp002 ~/bin $ ldd ~/bin/ribasim/bin/ribasim linux-vdso.so.1 (0x00007ffe803d2000) libdl.so.2 => /lib64/libdl.so.2 (0x00007f79b3905000) libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007f79b36ed000) libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f79b34cd000) libc.so.6 => /lib64/libc.so.6 (0x00007f79b30f7000) /lib64/ld-linux-x86-64.so.2 (0x00007f79b3b09000) visser_mn@v-h7rdp002 ~/bin $ ldd ~/bin/ribasim/lib/libribasim.so linux-vdso.so.1 (0x00007ffc99973000) libjulia.so.1.12 => /u/visser_mn/bin/ribasim/lib/libjulia.so.1.12 (0x00007fcc064d7000) libjulia-internal.so.1.12 => not found libc.so.6 => /lib64/libc.so.6 (0x00007fcbc1277000) libdl.so.2 => /lib64/libdl.so.2 (0x00007fcbc1073000) libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fcbc0e53000) /lib64/ld-linux-x86-64.so.2 (0x00007fcc062cd000) visser_mn@v-h7rdp002 ~/bin $ export LD_LIBRARY_PATH=~/bin/ribasim/lib:~/bin/ribasim/lib/julia:$LD_LIBRARY_PATH visser_mn@v-h7rdp002 ~/bin $ ribasim/bin/ribasim ~/ribasim-nl/trivial/ribasim.toml ┌ Info: Starting a Ribasim simulation at 2026-01-26T23:16:08.652. │ toml_path = "/u/visser_mn/ribasim-nl/trivial/ribasim.toml" │ cli.ribasim_version = "2026.1.0-rc1" │ starttime = 2020-01-01T00:00:00 │ endtime = 2021-01-01T00:00:00 └ threads = 1 ┌ Warning: The following experimental features are enabled: concentration └ @ Ribasim /opt/teamcityagent/work/ecd2b8f9b25b1609/ribasim/core/src/logging.jl:51 [ Info: Computation time: 503 milliseconds [ Info: The model finished successfully at 2026-01-26T23:16:14.037. ``` It seems the default of the JuliaC LinkRecipe is the magic string `@julia`, whereas we need the magic string `@bundle`. ``` $ objdump -p libribasim.so | grep -E 'RPATH|RUNPATH' RPATH /u/svc-teamcity-ansible/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/lib:/u/svc-teamcity-ansible/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/lib/julia ``` </p> </details>
Author
|
Bump :) |
Fixes JuliaLang#33. This requires Julia 1.12.3 or later.
Author
|
Is there still interest in this PR? |
visr
added a commit
to Deltares/Ribasim
that referenced
this pull request
May 29, 2026
I recently force-pushed JuliaLang/JuliaC.jl#47. But doing that breaks our Manifest.toml. So I made a separate branch meant for Ribasim instead. This also updates PythonCall and JuliaCall to the latest patch releases. These are pinned since the versions need to match between Pkg and Pixi.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fix #33.
I was hoping it was just a matter of changing the defaults, but it seems to cause segfaults for me running testset "Programmatic binary (trim)" locally on Windows.EDIT: indeed fixed by JuliaLang/julia#58952 (comment), so requires 1.12.3