multicoretests
multicoretests copied to clipboard
PBT testsuite and libraries for testing multicore OCaml
We should extend the test suite to include more modules from the Stdlib. There's an overview here: https://github.com/ocaml/ocaml/issues/10960#issue-1117254143 We may want to prioritize tests of modules with a `Y` in...
Today surfaced a Windows trunk bytecode crash on src/domain/domain_spawntree.ml https://github.com/ocaml-multicore/multicoretests/actions/runs/5154525696/jobs/9283085877 ``` random seed: 502502158 generated error fail pass / total time test name [ ] 0 0 0 0 /...
We came across a CI run on Linux trunk/5.2 ARM64 which timed out in `thread_createtree - with Atomic` despite none of the previous tests taking substantial time. https://ocaml-multicoretests.ci.dev:8100/job/2023-12-12/132004-ci-ocluster-build-19fadd ``` random...
On #411 we saw a segfault in `Lin Out_channel test with Domain` in a 32-bit trunk/5.2 run: https://github.com/ocaml-multicore/multicoretests/actions/runs/6826699760/job/18567181221?pr=411 ``` random seed: 125264582 generated error fail pass / total time test...
We are observing cases where the test suite runs slowly and ends up timing out on s390x after 300m / 5h. - Dec 5: https://ocaml-multicoretests.ci.dev:8100/job/2023-12-05/152822-ci-ocluster-build-ac98e8 - Dec 5: https://ocaml-multicoretests.ci.dev:8100/job/2023-12-05/222625-ci-ocluster-build-9be70b -...
Off-line input from @polytypic: Using `QCheck STM` to do a simple sequential test with `(implicit_transitive_deps false)` seems to require adding no less than four dependencies: `qcheck-core`, `qcheck-core.runner`, `qcheck-stm.stm`, `qcheck-stm.sequential`. To...
This is an existing issue which we have been observing and battling repeatedly. I now saw it again and decided to track it separately from the general #338 and #384....
In #409 a FreeBSD CI run on 5.0.0 failed to trigger a counterexample in `STM Sys test parallel` within 1000 attempts: https://ocaml-multicoretests.ci.dev:8100/job/2023-11-09/084832-ci-ocluster-build-542efa ``` random seed: 4002063 generated error fail pass...
We are starting to have a wonderful torture test suite. It is however only run on Github Actions machines: Intel Linux and Intel Mac in native code mode... Other torture...
Since adding `agree_prop_par_asym` to `STM_domain` in #315 we have seen occasional cases of the `_ ref test parallel asymmetric` failing to trigger. Here's a fresh case from yesterday, on Cygwin...