Samuel Thibault
|
0e79ad3c82
Add STARPU_PERF_MODEL_HOMOGENEOUS_CPU environment variable to allow having one perfmodel per CPU core
|
8 anni fa |
Samuel Thibault
|
603bf61baf
fix documentation about limit_cpu_mem
|
8 anni fa |
Samuel Thibault
|
951862353b
Add break on executing a given task
|
8 anni fa |
Luka Stanisic
|
70f4c2737c
doc: explain scheduling cost in simulation
|
8 anni fa |
Luka Stanisic
|
db4d8cf053
simgrid: taking into account scheduling cost
|
8 anni fa |
Luka Stanisic
|
3fab90c341
doc: correcting typo for PERF_MODEL_HOMOGENEOUS
|
8 anni fa |
Luka Stanisic
|
e310fc7224
simgrid: adding another environment variable STARPU_SIMGRID_FETCHING_INPUT_COST, to simulate the time to perform a fetching input
|
8 anni fa |
Samuel Thibault
|
7cdeadd447
Add STARPU_CUDA_THREAD_PER_DEV environment variable to support driving all
|
8 anni fa |
Samuel Thibault
|
5278ab7d1b
typo
|
8 anni fa |
Samuel Thibault
|
14492fb09a
typo
|
8 anni fa |
Samuel Thibault
|
72453f9f98
Disable forcing available memory by default, since this makes StarPU spuriously free data when prefetching fills the memory. Clean buffers are enough to be able to allocate data easily anyway.
|
8 anni fa |
Samuel Thibault
|
ffcd176304
Fix unit for STARPU_DISK_SWAP_SIZE: Bytes is really too small, use MB just like LIMIT*
|
8 anni fa |
Corentin Salingue
|
74882f89bc
add more pieces to the doc for the MPI master-slave
|
8 anni fa |
Corentin Salingue
|
5d3f3a018a
Update doc with MPI Master-Slave support + fix visibility of 2 sections in MPI doc
|
8 anni fa |
Nathalie Furmento
|
fd7acb2f0f
fix variable names which allow to share performance models between similar devices
|
8 anni fa |
Samuel Thibault
|
8cf58b4493
Allow to share performance models between CUDA devices
|
8 anni fa |
Samuel Thibault
|
37578a0d00
Add STARPU_MPI_FAKE_SIZE and STARPU_MPI_FAKE_RANK to allow simulating
|
8 anni fa |
Samuel Thibault
|
ad53cb20ee
Add experimental STARPU_DIDUSE_BARRIER environment variable support
|
8 anni fa |
Samuel Thibault
|
87de25ac25
revert spurious changes
|
8 anni fa |
Samuel Thibault
|
dde9701d14
follow-up r19535 with same fix below
|
8 anni fa |
Samuel Thibault
|
3ade010051
rename STARPU_ONE_THREAD_PER_STREAM to STARPU_CUDA_THREAD_PER_WORKER
|
8 anni fa |
Andra Hugo
|
4436cde002
forgot copyrights
|
8 anni fa |
Andra Hugo
|
15103b9723
GPU partitioning using contexts, and fixing dmda to consider the pipeline len
|
8 anni fa |
Samuel Thibault
|
c6dd318e3f
prevent starpu_tasks_rec_complete from overwriting the FxT trace
|
8 anni fa |
Luka Stanisic
|
7437946cf8
simgrid: adding environment variable to take into account task submission costs in simgrid mode
|
8 anni fa |
Samuel Thibault
|
cdbf49888a
Fix thread binding on knf in accelerator mode
|
8 anni fa |
Nathalie Furmento
|
911797a49a
doc: corrections
|
8 anni fa |
Nathalie Furmento
|
3bd9664c1c
doc: rename files to have a html documentation generated with the correct order
|
8 anni fa |