 Samuel Thibault
|
1f1954c43b
Add OpenCL kernel submission pipelining, to overlap costs
|
11 anni fa |
|
Samuel Thibault
|
f99ef13c88
Add CUDA kernel submission pipelining, to overlap costs and allow concurrent
|
11 anni fa |
|
Samuel Thibault
|
aa5c3e5c50
Let interfaces declare which transfers they allow with the can_copy methode.
|
11 anni fa |
|
Samuel Thibault
|
89ece6b14f
changelog
|
11 anni fa |
|
Samuel Thibault
|
2a1bb7bf49
Update 1.1.2 svn revision
|
11 anni fa |
|
Samuel Thibault
|
14b8f531a6
Lock performance model files while writing and reading them to avoid issues on parallel launches, MPI runs notably.
|
11 anni fa |
|
Samuel Thibault
|
f43bf86292
Fix complexity of implicit task/data dependency, from quadratic to linear
|
11 anni fa |
|
Samuel Thibault
|
3db86a2ede
* Add STARPU_VARIABLE_NBUFFERS to be set in cl.nbuffers, and nbuffers and
|
11 anni fa |
 Nathalie Furmento
|
53a11e7402
workers: new function starpu_worker_get_type_as_string()
|
11 anni fa |
|
Samuel Thibault
|
3b7a4ec8f9
Mention MIC too
|
11 anni fa |
|
Samuel Thibault
|
3f9b8cd37a
Add the starpu_paje_summary statistics tool.
|
11 anni fa |
|
Samuel Thibault
|
214c709b08
port r 13033 from 1.1: Add changelog
|
11 anni fa |
|
Samuel Thibault
|
8d8e0903d0
port r12948 from 1.1: Add changelog entry
|
11 anni fa |
|
Samuel Thibault
|
7dd6b42215
Document that blocking spinlocks are only available on Linux
|
11 anni fa |
|
Samuel Thibault
|
039524c7cf
Spinlocks now block after a hundred tries. This avoids typical 10ms pauses
|
11 anni fa |
|
Samuel Thibault
|
0c9c7c0ba3
advertise lws
|
11 anni fa |
|
Samuel Thibault
|
4705435be8
Add STARPU_TAG_ONLY, to specify a tag for traces without making StarPU manage the tag.
|
11 anni fa |
|
Samuel Thibault
|
778b4c444a
* Add codelet size, footprint and tag id in the paje trace.
|
11 anni fa |
|
Nathalie Furmento
|
b43914770b
ChangeLog: move updates to the correct section, some updates were wrongly listed as part of release 1.1.x instead of 1.2.x
|
11 anni fa |
|
Samuel Thibault
|
1d8e85f121
port r12828 from 1.1: Add changelog for STARPU_CALIBRATE_MINIMUM
|
11 anni fa |
|
Samuel Thibault
|
d592415088
port r12811 from 1.1: Add changelog for the "scheduling" state in traces
|
11 anni fa |
|
Nathalie Furmento
|
06778e0bb6
sched_ctx: new function starpu_sched_ctx_display_workers() to display the worker names of a context
|
11 anni fa |
|
Nathalie Furmento
|
90879cd44c
configure: allow to set --enable-verbose=extra to increase debug messages verbosity
|
11 anni fa |
|
Samuel Thibault
|
f255d2c95a
port r12703 from 1.1: Allow data to be uninitialized but the application provide an init_cl to initialize it
|
11 anni fa |
|
Nathalie Furmento
|
cdb8ad6182
ChangeLog: update
|
11 anni fa |
|
Nathalie Furmento
|
68d0971b70
starpufft: By default, StarPU FFT examples are not compiled and checked, the configure option --enable-starpufft-examples needs to be specified to change this behaviour.
|
11 anni fa |
|
Samuel Thibault
|
f2b1d0f595
port r12584 from 1.1: Re-enable asynchronous transfers from CPUs, now that the issue has been found to be in application code
|
11 anni fa |
|
Samuel Thibault
|
092f322b1c
Add CUDA concurrent kernel execution support through the STARPU_NWORKER_PER_CUDA environment variable.
|
11 anni fa |
|
Samuel Thibault
|
ea5d8f596b
Make a copy of the interface to the memchunk only when the latter gets detached from the data, and thus the interface code will not work on it. Drop the copy when the memchunk gets reattached. This allows interfaces to modify pointers in the interface in unpack, notably
|
11 anni fa |
|
Samuel Thibault
|
1fcd50fd76
Add changelog entry
|
11 anni fa |
