 Samuel Thibault
|
7ba4274ea3
drop spurious include
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
c7291e5cb2
Use _starpu_callback_func_t instead of explicit function type, which bothers MSVC
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
1f1954c43b
Add OpenCL kernel submission pipelining, to overlap costs
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
b7efebe04d
Fix task accounting when the scheduler erroneously puts a non-cuda task to a cuda device
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
3abd76c814
shutdown StarPU if it was successfully started but the devices we wanted are not there
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
86eab72d26
revert r13284 and r13286: probes happen before the TSD is allocated
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
df782f4617
delete TSD only after fxt is really off
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
bec10be1cd
Use TSD to cache the value of gettid for FXT traces
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
f313079113
Factorize code to prepare for OpenCL pipelining
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
62e0ff89af
Fix missing include for windows port of locking
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
23636fb4f4
Fix simgrid build
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
77e107e4af
Warn if the user uses concurrent execution on a device which doesn't support it
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
62194c5c86
Fix filling the pipeline
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
f99ef13c88
Add CUDA kernel submission pipelining, to overlap costs and allow concurrent
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
647ccc35a7
Fix estimated allocated memory
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
41cfdc75cd
free data outside critical section
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
3b766361ce
free data outside critical section
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
206b3adf49
port r13269 from 1.1: free data outside critical section
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
aa5c3e5c50
Let interfaces declare which transfers they allow with the can_copy methode.
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
1ec6838c01
Implement perfmodel locking on windows
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
e5cf638f86
Pass codelet model when computing footprints for the bound computation
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
b294c2f942
Fix hardcoded limitation to 64 MPI nodes in traces
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
f8ff9c7d1c
Fix build without memcpy_peer
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
89ece6b14f
changelog
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
2c6e5a75b0
enable gpu-gpu transfers for matrices
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
d76f89457d
Separate bitfields set by starpu from bitfields set by user, since the latter are accessed unsafely by starpu, but the former shouldn't
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
2a1bb7bf49
Update 1.1.2 svn revision
|
%!s(int64=11) %!d(string=hai) anos |
|
Samuel Thibault
|
14b8f531a6
Lock performance model files while writing and reading them to avoid issues on parallel launches, MPI runs notably.
|
%!s(int64=11) %!d(string=hai) anos |
 Nathalie Furmento
|
6d015be440
tests/main/{subgraph_repeat_tag.c,subgraph_repeat_regenerate_tag.c}: stop regenerating tasks when all loops are done, this permits to call starpu_data_unregister(), otherwise it would wait for the tasks
|
%!s(int64=11) %!d(string=hai) anos |
|
Nathalie Furmento
|
c232f22ddd
examples/spmv/dw_block_spmv.c: fix following #13213
|
%!s(int64=11) %!d(string=hai) anos |
