aboutsummaryrefslogtreecommitdiffhomepage
path: root/stm/asyncTaskQueue.mli
Commit message (Collapse)AuthorAge
* *Queue: API to wake up all threadsGravatar Enrico Tassi2015-02-16
|
* Update headers.Gravatar Maxime Dénès2015-01-12
|
* AsyncTaskQueue: simpler model (no parking area, continuation tasks)Gravatar Enrico Tassi2014-12-17
|
* AsyncTaskQueue: parsin can also happen in the workers nowGravatar Enrico Tassi2014-11-27
|
* AsyncTaskQueue: API to park a workerGravatar Enrico Tassi2014-11-27
| | | | | | | | | | | | Generalize the old model by letting one park a worker and by letting the (parked) worker be picky about the tasks it picks up. The use of that is the following: a proof worker, while performing its "main" task (building a proof term) computes all the intermediate states but returns only its main result. One can ask the worker to hang around, and react to special tasks, like printing the goals of an intermediate state.
* STM: code refactoringGravatar Enrico Tassi2014-11-03
| | | | | This is mainly shuffling code around and removing internal refs that are not needed anymore.
* STM: primitives to snapshot a .vi while in interactive modeGravatar Enrico Tassi2014-10-13
|
* STM: new "par:" goal selector, like "all:" but in parallelGravatar Enrico Tassi2014-08-05
| | | | | par: distributes the goals among a number of workers given by -async-proofs-tac-j (defaults to 2).
* STM: code restructured to reuse task queue for tacticsGravatar Enrico Tassi2014-08-05