Olexiy Kulchitskiy (Apr 27 2022 at 16:09): Olexiy Kulchitskiy (Apr 27 2022 at 16:09):Hi, guys! I'm new to all this WASM stuff and looking for deterministic WASM runtime (for some POC). Please advise whether Wasmtime is handling non-determinism issues and if yes, how (would be nice to point to some article/code)?
Chris Fallin (Apr 27 2022 at 17:31):Hi @Olexiy Kulchitskiy -- the execution of Wasm in Wasmtime should be completely deterministic with a few exceptions:
Chris Fallin (Apr 27 2022 at 17:32):Ah, stack depth as well; the point at which one gets a stack-overflow trap is implementation-defined (but that should not be observable in any normally-returning Wasm call, as the trap terminates the instance)
bjorn3 (Apr 27 2022 at 17:56):And the memory allocation limit.
Dan Gohman (Apr 27 2022 at 18:03):You may also want to use the static_memory_forced API, to prevent memory.grow allocation from failing nondeterministically.
Olexiy Kulchitskiy (Apr 27 2022 at 18:57):got it. Thank you
Lars Hansen (May 02 2022 at 08:44):Relaxed SIMD will be deterministic per-platform but will appear nondeterministic across platforms - basically the same problem as for NaNs, but will require a different solution.
dakom (Mar 26 2026 at 14:16):Resurrecting this after a while, could someone expand on this for wasi-p2, and maybe even wasi-p3 with async in mind?
Joel Dice (Mar 26 2026 at 14:37):You could start by doing a text search for "determinstic" in https://github.com/WebAssembly/component-model/blob/main/design/mvp/CanonicalABI.md. Async task and thread scheduling are indeed nondeterministic, spec-wise, in the default profile, but enabling the deterministic profile can reduce (and maybe eliminate?) that nondeterminism.
fitzgen (he/him) (Mar 26 2026 at 15:43):@dakom have you read https://docs.wasmtime.dev/examples-deterministic-wasm-execution.html already? it attempts to answer your original question, if anything is missing we should update that page to contain it
dakom (Mar 26 2026 at 17:37):@fitzgen (he/him) - that's an excellent resource, thank you!!
So I think my only high-level question is how this page might need to be updated for p3/async?
fitzgen (he/him) (Mar 26 2026 at 17:39):that all falls under "imports" -- if you want determinism, you need to make sure your imports are deterministic (or you are deterministically replaying a recording of non-deterministic imports) and http APIs are not particularly deterministic
Joel Dice (Mar 26 2026 at 17:48):It partially falls under "imports", but some of the "imports" for async are intrinsics over which the embedder has no control, and Wasmtime's implementation of some of those intrinsics are non-deterministic, with no option (yet) to make them deterministic.
dakom (Mar 26 2026 at 17:49):It would also depend on the runtime being run in a deterministic way, e.g. tokio, right?
fitzgen (he/him) (Mar 26 2026 at 17:49):ah I would have assumed that like task scheduling was deterministic assuming that the imports themselves were deterministic
Joel Dice (Mar 26 2026 at 17:51):I guess I would say we haven't made an effort to make task scheduling deterministic yet, so if it is, that's a happy coincidence :)
Joel Dice (Mar 26 2026 at 17:54):@dakom FYI Wasmtime's async support doesn't depend on tokio (although wasmtime-wasi does), so you could bring your own deterministic executor if desired.
Joel Dice (Mar 26 2026 at 17:55):But I suspect there are sources of nondeterminism inside Wasmtime's scheduling even with a deterministic executor, just because we haven't tried to avoid that yet.
Last updated: Apr 12 2026 at 23:10 UTC