 Samuel Thibault
|
804e6f8233
properly test for task termination
|
11 lat temu |
 Marc Sergent
|
0026768555
Adding the handling of multiple fathers in hierarchical schedulers,
|
11 lat temu |
|
Marc Sergent
|
e998e008c6
Merge from trunk @11714:11788
|
11 lat temu |
 Andra Hugo
|
2769aaa8cb
hypervisor: correct computation of idle
|
11 lat temu |
|
Samuel Thibault
|
4d81325ef7
fix missing initialization
|
11 lat temu |
|
Samuel Thibault
|
b8062a9fba
keep pushed back tasks in proper order
|
12 lat temu |
|
Andra Hugo
|
20906d8d77
hyp only: optimise resize (use idle and speed for the triggering policy)
|
12 lat temu |
|
Andra Hugo
|
6266721e89
do not take the mutex if the current worker is calling the code
|
12 lat temu |
|
Andra Hugo
|
39b06ba608
do not take changing_ctx_mutex when finishing executing the tasks (in decrement_nsubmitted_tasks), just check if the counter is 1, do what you have to do, and then decrement it (decrement the counter while verifying it wakes up the the thread calling wait_for_all)
|
12 lat temu |
|
Marc Sergent
|
fe685563e7
Adding several features which needs each other to compute correctly :
|
12 lat temu |
|
Samuel Thibault
|
b70f84e342
Properly find the task which optimizes computation time benefit
|
12 lat temu |
|
Samuel Thibault
|
a62e0dba32
better comment
|
12 lat temu |
|
Samuel Thibault
|
a06ea29938
Use separate expected ends, to avoid systematically preferring small tasks, possibly leading to starvation
|
12 lat temu |
|
Samuel Thibault
|
4f82d35ce9
Avoid keeping the lock held while scheduling tasks, to let scheduling happen in parallel. This means several processors might take 5 tasks each, thus emptying the fifo, but at least we don't have a huge contention point
|
12 lat temu |
|
Samuel Thibault
|
23b0ea64b9
Add a heft variant which draws tasks from its own fifo
|
12 lat temu |
|
Samuel Thibault
|
30c9f2732c
Plug perfmodel_select children normally from the scheduling strategy, instead of internally within perfmodel_select. This makes the room call go through the MCT node
|
12 lat temu |
|
Samuel Thibault
|
f7684ba8e2
keep the original data transfer time for the beta penalty, otherwise it seems we do not take communications into account enough to achieve overlapping. Proper communication scheduling would allow to replace this heuristic
|
12 lat temu |
|
Samuel Thibault
|
b693f565e0
Handle case when there is no worker any more for a given scheduling node
|
12 lat temu |
|
Samuel Thibault
|
ae4b92a838
Add missing files
|
12 lat temu |
|
Samuel Thibault
|
11266adce9
FIXME
|
12 lat temu |
|
Samuel Thibault
|
2441190259
factorize MCT code
|
12 lat temu |
|
Samuel Thibault
|
231fa5c8d4
fix spacing
|
12 lat temu |
|
Samuel Thibault
|
2240d43c8f
more cleanup
|
12 lat temu |
|
Samuel Thibault
|
716c571d26
clean up MCT computations
|
12 lat temu |
|
Marc Sergent
|
7a983bdaa4
Correcting copyrights on src/sched_policies/
|
12 lat temu |
 Nathalie Furmento
|
0d7c8ed948
merge trunk@9680, trunk@9683: tests/sched_policies/prio.c: use portable random function, drand48 returns values between 0 and 1
|
12 lat temu |
|
Marc Sergent
|
496771f13c
src/sched_policies/node_worker.c : Bug fix concerning the way to get a combined worker's id.
|
12 lat temu |
|
Marc Sergent
|
db2bcd07f6
Merge from trunk @11646:11713
|
12 lat temu |
|
Nathalie Furmento
|
ed41317844
examples/pi/SobolQRNG/sobol_gold.c: include strings.h
|
12 lat temu |
|
Marc Sergent
|
76691542fa
Adding a prefetch field in the task' structure to optimize unneeded prefetching calls, and adapting default threshold values for some predefined hierarchical schedulers.
|
12 lat temu |
