 Cédric Augonnet
|
554cb8bc74
dynamically load an empty kernel for the microbench that measure overhead
|
17 lat temu |
|
Cédric Augonnet
|
76b022ddf0
do not forget to include ppm-downscaler.h in the distribw
|
17 lat temu |
|
Cédric Augonnet
|
c5e951e16a
we need to get rid of .sampling/* between every iteration, not only at the beginning !
|
17 lat temu |
|
Cédric Augonnet
|
be54ec551f
rework the gnuplot output of the "error-model" test
|
17 lat temu |
|
Cédric Augonnet
|
3521786f90
add a *terribly* ugly script to measure the accuracy of performance prediction.
|
17 lat temu |
|
Cédric Augonnet
|
95f9c1a844
When the "verbose" option is enabled, create a log file. Its content is filled
|
17 lat temu |
|
Cédric Augonnet
|
c59fcbc747
do not display debug information after pthread_exit
|
17 lat temu |
|
Cédric Augonnet
|
72528f57d7
Measure the difference between predicted execution time and measured execution
|
17 lat temu |
|
Cédric Augonnet
|
01ed5d79fd
- use a performance models based on history
|
17 lat temu |
|
Cédric Augonnet
|
d5f7fb9d95
add a new benchmark to measure the performance of sgemm on square matrices (to
|
17 lat temu |
 Samuel Thibault
|
153901a43e
typos
|
17 lat temu |
|
Cédric Augonnet
|
81375feb05
add a new example of matrix multiplication with no filter, only one BLAS ops is
|
17 lat temu |
|
Cédric Augonnet
|
91bc53fb3e
Include the worker identifier in the performance model debug logs.
|
17 lat temu |
|
Cédric Augonnet
|
a729df228a
it is possible to specify where to bind the different worker using the
|
17 lat temu |
|
Cédric Augonnet
|
8a02f335ee
underline that task submission is asynchronous
|
17 lat temu |
|
Cédric Augonnet
|
c37b910143
document the basic examples so that they should be self-sufficient
|
17 lat temu |
|
Cédric Augonnet
|
232d016ed7
implement a very simple matrix multiplication that uses filters
|
17 lat temu |
|
Cédric Augonnet
|
e6c184c3dc
specify what is described/explained in the hello-world example
|
17 lat temu |
|
Cédric Augonnet
|
1d8b7066e9
- put a short description of the goal of the vector-scal example
|
17 lat temu |
|
Cédric Augonnet
|
3245db6395
starts to document the "vector-scal" example
|
17 lat temu |
|
Cédric Augonnet
|
8a17554361
Add a trivial example to show how to scale a vector by a constant factor.
|
17 lat temu |
|
Cédric Augonnet
|
5eb2d8a236
correct an error in the documentation
|
17 lat temu |
|
Cédric Augonnet
|
4ed11473ae
Comment the "Hello World" example to make it self sufficient (hopefully).
|
17 lat temu |
|
Cédric Augonnet
|
df3e02ea72
remove a deprecated warning
|
17 lat temu |
|
Cédric Augonnet
|
6f1c146e07
In case CPUs' clocks are not synchronized, we cannot use ticks so we use a
|
17 lat temu |
|
Cédric Augonnet
|
366cec56e8
pthread_spinlock_t have no static initialization value, so we need to make sure
|
17 lat temu |
|
Cédric Augonnet
|
d84888c8f9
add timing in the YUV downscaler
|
17 lat temu |
|
Cédric Augonnet
|
846e608648
use StarPU to perform YUV downscaling (CPU only yet)
|
17 lat temu |
|
Cédric Augonnet
|
bc3727d2f0
add a new example, to downscale a YUV movie (nothing yet ..)
|
17 lat temu |
|
Cédric Augonnet
|
626b22ba6f
introduce a new example "downscaler", no real stuff for the moment
|
17 lat temu |
