Repository navigation
Exclude the JAX output stage's compilation from the VMEC++ runtime - #12
Merged
Merged
Conversation
Since vmecpp#861 the first vmecpp.run of a process at a given resolution compiles the JAX output stage, about a second: the dshape case then took 1.34 s against 0.06 s for VMEC2000 and failed the runtime check. An untimed one-iteration run at the final resolution triggers the compilation before the timed run; its runtime is logged.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Since vmecpp#861, the first
vmecpp.runof a process at a given resolution compiles the JAX output stage (about 1 s). The V&V run on vmecpp main then failed the runtime check: dshape took 1.34 s against 0.06 s for VMEC2000 (https://github.com/proximafusion/vmecpp/actions/runs/36309334086/job/108592125248).run_vmecppnow makes an untimed one-iteration run at the configuration's final radial resolution before the timed run. The compilation depends only on the grid sizes, so this triggers it; the warm-up runtime is written to the run's log. The timed run and the saved wout are unchanged.Local timings (warm-up / timed): solovev 0.64 s / 0.025 s, cma 0.68 s / 0.25 s, cth_like_fixed_bdy 0.66 s / 0.10 s; a cold solovev run takes 0.63 s. The full suite was not run locally (it needs VMEC2000 in docker).