Timeline for DistributeDefinitions and synchronization in Mathematica 10
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 7, 2016 at 6:47 | comment | added | Eisbär | Thats a tricky one, as it is not actually returning the true value as given by the parallel kernel but an interpretation of that by the master kernel. I have to rember that one. Thank you for elaborating. | |
| Jun 5, 2016 at 6:50 | comment | added | Fred Simons | @Eisbär. After ParallelEvaluate[Clear[z]], the value of z in the subkernels is cleared, but in the main kernel the valueof z is still 4. Therefore, ParallelEvaluate[z] (which now does not redistribute the definition of z to the subkernels) returns the list {z,z,z,z} in the main kernel. But in the main kernel this list further evaluates to {4,4,4,4}. | |
| May 25, 2016 at 10:35 | comment | added | Eisbär | I still do not get why we get ParallelEvaluate[z] (* {4,4,4,4} *) after the Clear[z] and still ValueQ[z] is false. Which mechanism makes the 4 know to ParallelEvaluate[z] after ParallelEvaluate[Clear[z]] if not DistributeDefinitions? | |
| May 25, 2016 at 9:28 | history | edited | Szabolcs | CC BY-SA 3.0 | added 98 characters in body |
| May 25, 2016 at 8:55 | history | edited | Szabolcs | CC BY-SA 3.0 | added 4 characters in body |
| Apr 6, 2016 at 18:03 | vote | accept | Fred Simons | ||
| Apr 6, 2016 at 18:01 | history | answered | Fred Simons | CC BY-SA 3.0 |