 Nathalie Furmento
|
61e6732e89
src/core/perfmodel/perfmodel_history.c: correctly index array model->nimpls and dump into model file the index of the global combinations array (i.e the internal arch_combs array)
|
лет назад: 12 |
|
Nathalie Furmento
|
552cbdbcf0
src/core:
|
лет назад: 12 |
|
Nathalie Furmento
|
892665a9b3
src/drivers/driver_common/driver_common.c: when dealing with parallel contexts, also generate trace and stats for non-master workers
|
лет назад: 12 |
|
Nathalie Furmento
|
c506a5901d
src/core/sched_ctx: new internal function _starpu_sched_ctx_get_sched_ctx_for_worker_and_job to return a context if the worker is the master of a parallel context, and the job is meant to be executed on this parallel context
|
лет назад: 12 |
 Andra Hugo
|
43402fb9b1
fix wrong allocation of the table of added workers
|
лет назад: 12 |
|
Andra Hugo
|
a7288fd535
build perf arch for contexts
|
лет назад: 12 |
|
Andra Hugo
|
694211e31c
fix debug mode compilation issues
|
лет назад: 12 |
|
Nathalie Furmento
|
8e6e6dcaab
revert Makefile.am to build tools and disable tools/starpu_perfmodel_plot.c
|
лет назад: 12 |
|
Andra Hugo
|
d795ede306
merge with trunk 13004
|
лет назад: 12 |
|
Andra Hugo
|
71cc7e4997
don't indicate number of impls they will be computed when executed
|
лет назад: 12 |
|
Andra Hugo
|
10fe309eac
compil issues
|
лет назад: 12 |
|
Nathalie Furmento
|
436a56e106
configure.ac: --enable-blas-lib now also accepts the value mkl (to be consistent with the other blas libraries)
|
лет назад: 12 |
|
Andra Hugo
|
c873b63e88
small fix
|
лет назад: 12 |
|
Andra Hugo
|
a9b35bd4a7
fix malloc & memset perf model
|
лет назад: 12 |
|
Andra Hugo
|
a0f55874a4
fixes for valid model
|
лет назад: 12 |
|
Andra Hugo
|
002b53889f
fix bugs with the ids of the arch combinations
|
лет назад: 12 |
 Samuel Thibault
|
02ce5c2683
Drop spurious changes (again...)
|
лет назад: 12 |
|
Andra Hugo
|
fca611cf90
small compil issue
|
лет назад: 12 |
|
Samuel Thibault
|
fe9ca6b56f
port r12993 from 1.1: Fix crash with a lot of MPI nodes
|
лет назад: 12 |
|
Andra Hugo
|
5593439277
changing perf_model structure: arch now contains several devices, so we can have for eg one device STARPU_CPU with 2 cores one device STARPU_CUDA_WORKER with 1 core
|
лет назад: 12 |
|
Andra Hugo
|
2eaf836156
merge with trunk
|
лет назад: 12 |
|
Samuel Thibault
|
f14e68dc2c
Separate worker and thread state, to see streamed kernel executions
|
лет назад: 12 |
|
Samuel Thibault
|
02acdca9b3
drop spurious changes
|
лет назад: 12 |
|
Samuel Thibault
|
e22c0d9891
Fix function name: the existing function deals with a thread, not a worker
|
лет назад: 12 |
|
Samuel Thibault
|
3c1e8b4b7f
Do not register threads several times
|
лет назад: 12 |
|
Samuel Thibault
|
9bd9a1d234
Add missing pre_exec hook for dm
|
лет назад: 12 |
|
Samuel Thibault
|
35d4d64183
Do not take the first measurement into account, it is very often quite bogus
|
лет назад: 12 |
|
Samuel Thibault
|
8d8e0903d0
port r12948 from 1.1: Add changelog entry
|
лет назад: 12 |
|
Samuel Thibault
|
b98077fba7
Add STARPU_TRACE_BUFFER_SIZE environment variable to easily set the FxT buffer size
|
лет назад: 12 |
|
Samuel Thibault
|
9218801246
Do not expose a known-to-be-racy-but-we-re-fine optimization to helgrind
|
лет назад: 12 |
