 Andra Hugo
|
39ccc67d88
add hypervisor strategy monitoring perf counter (not portable, the counter reading the flops dependent on the architecture)
|
9 年之前 |
|
Andra Hugo
|
ba3e4d7403
small fixes done while writing my thesis + task window at context level (before pushing tasks to the scheduler) expressed w.r.t. to time and not number of tasks
|
10 年之前 |
|
Andra Hugo
|
5f388d2a95
hierarchical contexts - bug fixing and DEBUG macro
|
11 年之前 |
|
Andra Hugo
|
491a6c6124
fix hierarchical resizing of contexts
|
11 年之前 |
|
Andra Hugo
|
60d44eb737
resize hierarchic ctxs hierarchically :)
|
11 年之前 |
 Nathalie Furmento
|
e97679a558
fxt: rename function starpu_trace_user_event() as starpu_fxt_trace_user_event()
|
11 年之前 |
|
Andra Hugo
|
bfb858fe1d
Add hierarchical contexts
|
11 年之前 |
|
Andra Hugo
|
20906d8d77
hyp only: optimise resize (use idle and speed for the triggering policy)
|
11 年之前 |
|
Andra Hugo
|
12dec4c5cc
hypervisor optimisation (avoid unnecessary notifications/locks/trylocks) + add the hyp the traces
|
11 年之前 |
|
Andra Hugo
|
5a22d83b5c
hypervisor: feft_lp: compute max by using the execution of the last sample instead of the idle time
|
11 年之前 |
|
Andra Hugo
|
1289532233
whenever there are a few resources the lp cannot assign, all contexts share these resources
|
11 年之前 |
|
Andra Hugo
|
9502f53f25
introduce min max as a constraint for the linear program
|
11 年之前 |
|
Andra Hugo
|
9f20f53ce9
fix size_ctxs (do not consider ready tasks for size, only for resize), speed is computed on fixed value of the time
|
11 年之前 |
|
Andra Hugo
|
339104ed55
for feft_lp distribute resources at each resize instead of moving idle workers from one ctx to another
|
11 年之前 |
|
Andra Hugo
|
abebb0c602
cleaning ispeed_lp code + free hwloc_worker_set when deleting the context
|
11 年之前 |
|
Andra Hugo
|
e66dd53eab
cleaning + better managing the types of workers
|
12 年之前 |
|
Andra Hugo
|
b0b9e3f308
small fix to resize ctxs when only cpus are available
|
12 年之前 |
|
Andra Hugo
|
054afc9556
allow calling the resizing linear programs from the appl
|
12 年之前 |
|
Andra Hugo
|
45c4f38b10
rename velocity -> speed
|
12 年之前 |
|
Andra Hugo
|
5fd5c871e6
too expensive the lock in copmpute_all_performance_prediction, exp_start is update in post_exec/pre_exec/etc also
|
12 年之前 |
|
Andra Hugo
|
8d8d2aa2f2
fix speed computation in the hypervisor
|
12 年之前 |
|
Andra Hugo
|
76282f234b
fixes for triggering the resize
|
12 年之前 |
|
Andra Hugo
|
d0b859619b
fixes + refact in the hyp
|
12 年之前 |
|
Andra Hugo
|
6bcd0d3c6e
rename + comments in the public interface
|
12 年之前 |
|
Andra Hugo
|
214d5d3628
renames public lp functions
|
12 年之前 |
|
Andra Hugo
|
545f5f5b3b
rename hyp last step
|
12 年之前 |
|
Andra Hugo
|
a1fffeb7c8
rename hyp step1
|
12 年之前 |
