 Andra Hugo
|
6a484de2b1
merge first step
|
13 年之前 |
|
Andra Hugo
|
125960a47b
rollback to r7157
|
13 年之前 |
 Nathalie Furmento
|
59cbfcd5f3
merge trunk@7104:7182
|
13 年之前 |
|
Andra Hugo
|
fedc28d7d3
merge examples
|
13 年之前 |
|
Andra Hugo
|
ec83e69660
|
14 年之前 |
 Samuel Thibault
|
0d9f08ae8d
Fix copyrights, thanks Ludo!
|
14 年之前 |
 Cédric Augonnet
|
9070fdaebc
Add missing headers
|
15 年之前 |
|
Cédric Augonnet
|
3f9d1a9977
Rename the write-back mechanism into write-through mechanism.
|
15 年之前 |
|
Nathalie Furmento
|
4a983beb04
Update copyright
|
15 年之前 |
|
Cédric Augonnet
|
2cc17737f6
Rename the functions that permit to access the memory managed by StarPU from
|
15 年之前 |
|
Nathalie Furmento
|
1df7add367
Data Interface API: rename macro to get fields of interfaces
|
15 年之前 |
|
Cédric Augonnet
|
aae569a5c8
- Rename "struct starpu_filter_t" into "struct starpu_data_filter" and remove
|
15 年之前 |
 Mehdi Juhoor
|
76127800bc
the starpu_filter structure now contains a nchildren field. This help resolving an ambiguity with the get_nchildren function, which now serves to compute non-trivial number of children. As a result, filter_arg is now used as an additional parameter in some specific filters, e.g starpu_divide_in_2_filter_func.
|
15 年之前 |
|
Mehdi Juhoor
|
7902862bca
Modify the filters. filter_func now only updates a child interface using the father interface. Additionally, the user may provide two functions to a filter: get_nchildren returns the wanted number of children. If NULL, filter_arg is used. get_child_ops enables the programmer to specify each child's interface. If NULL, the father's interface is used.
|
15 年之前 |
|
Cédric Augonnet
|
cde9cbf959
Fix a bad use of the engineering notation: 10e9 -> 1e9
|
15 年之前 |
|
Samuel Thibault
|
c2a62115aa
fix windows cuda build, disable opencl on windows for now (no fork there...)
|
15 年之前 |
|
Nathalie Furmento
|
f58ca0920e
Renaming of public StarPU functions to provide a more consistent API.
|
15 年之前 |
|
Nathalie Furmento
|
b96d3703df
rename starpu_submit_task into starpu_task_submit
|
15 年之前 |
|
Cédric Augonnet
|
c037406638
Make it possible to select the number of tasks at runtime with the "-ntasks"
|
15 年之前 |
|
Cédric Augonnet
|
feff8b1018
Also display the distribution of the tasks between the different processing units
|
15 年之前 |
|
Cédric Augonnet
|
be0570c9d2
- Use a write-back cache policy for the "counter" data
|
15 年之前 |
|
Cédric Augonnet
|
6b2bcd48a5
- Rather than generating the random numbers on the host (which introduces a
|
15 年之前 |
|
Cédric Augonnet
|
8fd98c847d
provide a performance model name for the pi example
|
15 年之前 |
|
Cédric Augonnet
|
d6d00f1f8b
Fix compilation if CUDA is not available
|
15 年之前 |
|
Cédric Augonnet
|
76f2fd6053
Implement a (really naive) CUDA kernel that computes pi with a Monte Carlo
|
15 年之前 |
|
Cédric Augonnet
|
7b30fad879
Add a very simple monte carlo example to compute pi (CPU only yet)
|
15 年之前 |
