 Cédric Augonnet
|
e344b2b6cc
This is a first step toward the merge of the data replicatate and mem_chunk
|
15 years ago |
 Samuel Thibault
|
c3d288d15b
reduce size of mult coverage test to make it fit in 1GB memory
|
15 years ago |
|
Samuel Thibault
|
6c64f8620b
avoid duplicate tests that don't pass with no more than 1GB memory
|
15 years ago |
|
Samuel Thibault
|
7dec4b91b6
reduce default cholesky size so it can fit on 1GB laptops
|
15 years ago |
|
Samuel Thibault
|
1c8dd1f59b
test coverage after all programs
|
15 years ago |
|
Samuel Thibault
|
4d4ff33f6c
TODO
|
15 years ago |
|
Samuel Thibault
|
67bca51106
make use of _starpu_worker_may_execute_task instead of hardcoding it
|
15 years ago |
|
Cédric Augonnet
|
e49bbc7b64
Minimalist documentation of headers in src/datawizard/
|
15 years ago |
|
Samuel Thibault
|
60caa1c253
make _starpu_worker_may_execute_task take a task instead of a where
|
15 years ago |
|
Cédric Augonnet
|
ec828bcdd3
Minimalist documentation for the headers in src/core/
|
15 years ago |
|
Samuel Thibault
|
d023ed950e
push decision of "where" parameter of task pops down to scheduling policies, turn it into a worker id, and propagate to pop_task functions too
|
15 years ago |
|
Cédric Augonnet
|
9491fa3d4e
Minimalist documentation of some internal functions
|
15 years ago |
|
Cédric Augonnet
|
118472e5a7
starpu_fetch_data and starpu_release_data now take data replicates
|
15 years ago |
 Nathalie Furmento
|
5086fd5fc4
doc: minor fixes
|
15 years ago |
|
Cédric Augonnet
|
c21c9078b4
Fix typo
|
15 years ago |
|
Nathalie Furmento
|
ff58f98ac5
doc: minor fixes
|
15 years ago |
|
Cédric Augonnet
|
ddb5759a50
Minor code cleanups
|
15 years ago |
|
Cédric Augonnet
|
16683d829a
Factorize code
|
15 years ago |
|
Nathalie Furmento
|
34ce26600e
examples/coverage/coverage.sh: force to use the file in the current directory
|
15 years ago |
|
Cédric Augonnet
|
1c3235beac
Bug fix: we put a copy of the saved interface into the replicate, there should
|
15 years ago |
|
Cédric Augonnet
|
fc5da2fa67
_starpu_allocate_memory_on_node now manipulates a data replicate rather than a
|
15 years ago |
|
Cédric Augonnet
|
f57828e9d1
Don't try to allocate a piece of data that is already allocated.
|
15 years ago |
|
Cédric Augonnet
|
f93912c825
Add a reference to the API of the performance feedback section.
|
15 years ago |
|
Cédric Augonnet
|
9539f9903c
Factorize some code
|
15 years ago |
|
Cédric Augonnet
|
5c353dbcb8
Minor cleanups
|
15 years ago |
|
Nathalie Furmento
|
553f253990
src: debug mechanism to log in and out functions calls (was helpful to debug MPI...)
|
15 years ago |
|
Nathalie Furmento
|
2e546731a4
src/core/workers.c: braces
|
15 years ago |
|
Nathalie Furmento
|
283f137057
src/core/jobs.c: _starpu_get_model_name(j) return NULL if j=NULL
|
15 years ago |
|
Nathalie Furmento
|
b5b4e9b9b1
mpi/tests/ring_async_implicit.c: setting the token to 0 before every recv is wrong
|
15 years ago |
|
Nathalie Furmento
|
5949ffe780
MPI: debug code
|
15 years ago |
