 Corentin Salingue
|
f45e0a0e7c
merge from trunk
|
8 years ago |
|
Corentin Salingue
|
6bcb21c3f1
Progress also data in NUMA nodes
|
8 years ago |
 Samuel Thibault
|
d6d6378102
Don't stay stuck for too long pushing requests to GPUs
|
8 years ago |
|
Samuel Thibault
|
922ff28a1e
Make datawizard_progress use worker 0 of worker sets as reference for determining which memory nodes to make progress
|
9 years ago |
|
Samuel Thibault
|
cb9e2e2495
Avoid static analysis warnings
|
9 years ago |
|
Samuel Thibault
|
9a4011b6a0
Fix _starpu_datawizard_progress calls from the application context
|
9 years ago |
|
Corentin Salingue
|
cc4eb6b654
fix indent
|
9 years ago |
|
Corentin Salingue
|
e1f46dab74
really fix r19261 : add prefix _starpu_ to worker_drives_memory
|
9 years ago |
|
Corentin Salingue
|
1361c43dbd
Inline new functions for managing memoy nodes
|
9 years ago |
|
Corentin Salingue
|
3bbc1443c4
Each worker can manage the good set of memories
|
9 years ago |
|
Samuel Thibault
|
0b5135e60e
re-apply r18403 again, we'll be using maxmin/precision simgrid parameter to coarsen the time
|
9 years ago |
|
Samuel Thibault
|
514f24d352
revert r18403 for now, it seems to increase starpu-mpi-simgrid execution time a lot
|
9 years ago |
|
Samuel Thibault
|
4fb4ce92aa
Reduce the granularity of simulation of CUDA and MPI probes, it is not that expensive.
|
9 years ago |
|
Samuel Thibault
|
6bd3bb1b96
revert spurious change
|
9 years ago |
|
Samuel Thibault
|
8bb9a66912
Fix data migration: we need to flush the cache too. Add starpu_mpi_data_migrate which does things properly
|
9 years ago |
|
Samuel Thibault
|
1683d35a0d
Fix msg.h inclusion according to new location
|
11 years ago |
 Nathalie Furmento
|
bb61f3808d
update cnrs copyright
|
11 years ago |
|
Samuel Thibault
|
4c6de74837
Add starpu_data_idle_prefetch_on_node
|
11 years ago |
|
Samuel Thibault
|
f433b48230
Add STARPU_SIMGRID_CUDA_MALLOC_COST and STARPU_SIMGRID_QUEUE_MALLOC_COST environment variables to control whether simgrid simulates CUDA costs or not
|
11 years ago |
|
Samuel Thibault
|
5a20717d13
periodically perform reclaiming if needed
|
11 years ago |
|
Samuel Thibault
|
3937eb658e
Make _starpu_memchunk_tidy available to the application
|
11 years ago |
|
Samuel Thibault
|
d737e2d729
Periodically reclaim memory instead of waiting the last moment.
|
11 years ago |
|
Samuel Thibault
|
2a197a52b7
This was fixed
|
11 years ago |
|
Samuel Thibault
|
4b1a16edde
Fix University name after the fusion..
|
11 years ago |
|
Samuel Thibault
|
7b91d914cd
Also push requests if some previous requests have terminated
|
11 years ago |
|
Samuel Thibault
|
38fbca49e9
Fix _starpu_datawizard_progress prototype: it does not actually return anything
|
12 years ago |
|
Samuel Thibault
|
7ce4ced85d
Let the CUDA driver progress while the GPU is computing
|
12 years ago |
|
Samuel Thibault
|
e9c70aac17
port r11376 from 1.1: Introduce rwlock_tryrdlock and trywrlock variants. Use them in the progression hooks to avoid getting stuck due to idling workers.
|
12 years ago |
|
Samuel Thibault
|
d3065fdc1b
Make valgrind yield to other threads when busy-looping, otherwise we may get stuck
|
13 years ago |
|
Nathalie Furmento
|
103e97960d
src/datawizard/datawizard.c: include .h file for function definition
|
13 years ago |
