Samuel Li
093db83474
almost pass all tests, move to Alaska to debug one more case
2016-09-16 19:20:53 -06:00
Samuel Li
32fb705c2c
sync test program
2016-09-13 19:53:17 -07:00
Samuel Li
a045a767fa
sync to use gdb
2016-09-13 13:10:38 -06:00
Samuel Li
7e756e3a5a
move to linux to debug
2016-09-11 16:14:01 -06:00
Samuel Li
a196e9b42d
timer tweak
2016-09-06 15:54:12 -07:00
Samuel Li
a07e6ab71e
fix a bug, run final tests
2016-09-06 14:28:46 -07:00
Samuel Li
42a80662d8
finish 2D re-write. Some tests pass and some others fail. Need to keep debugging
2016-09-05 19:52:35 -06:00
Samuel Li
02cede6937
need to change IDWTHelper API a little bit
2016-09-01 18:47:50 -06:00
Samuel Li
c4387bb71c
need to debug a memory problem
2016-08-31 16:03:25 -06:00
Samuel Li
858895d3c8
Extend2D passes initial test, need to test more.
2016-08-26 14:51:23 -06:00
Samuel Li
4bda2f3042
one left extend worklet works correct
2016-08-26 10:54:47 -06:00
Samuel Li
87e80fc9c0
Merge branch 'master' into wavelet2
2016-08-24 15:53:52 -06:00
Samuel Li
f9dc98d0e6
add timers to report worklet computation time
2016-08-23 17:52:40 -06:00
Samuel Li
c5613b2cbb
optimized a few worklets
2016-08-23 08:56:03 -07:00
Samuel Li
ec185e3ade
WaveletFilter now stores array handles, so less data transfer between host and device
2016-08-22 14:28:29 -07:00
Samuel Li
daa80669f2
get rid of many small memory transfers
2016-08-22 13:49:13 -07:00
Samuel Li
553e396c6e
2D decompose/reconstruct seem to work well
2016-08-22 10:13:45 -06:00
Robert Maynard
ce3c7f64b5
Merge topic 'clip_restructure'
...
af05ea0e Cleanup clipping so we can use it from a filter.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !508
2016-08-22 09:56:21 -04:00
Samuel Li
165199c890
keep fighting on compiler errors
2016-08-21 23:10:05 -06:00
Samuel Li
372cd6729b
merge updates from the 1D wavelets branch
2016-08-18 22:42:29 -06:00
Robert Maynard
9712de8b24
Merge topic 'wavelets'
...
d677d0d1 small tweaks
816364d2 in an effort to get rid of a warning
778da350 In attempt to fix errors and warnings
bb450c51 fix a warning
49e56b61 two new wavelet filters, HAAR and CDF8/4 supported now
767356bc working on even length filters; need ASYM* support in Extend1D()
a6efad04 half done even length filters implementation
ee32ea4c took off timing code
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !482
2016-08-18 15:47:33 -04:00
Robert Maynard
af05ea0e8d
Cleanup clipping so we can use it from a filter.
2016-08-18 12:50:20 -04:00
Samuel Li
49e56b61b5
two new wavelet filters, HAAR and CDF8/4 supported now
2016-08-16 15:20:19 -06:00
Samuel Li
767356bc72
working on even length filters; need ASYM* support in Extend1D()
2016-08-15 17:54:15 -06:00
Samuel Li
3a1cfe5c6a
2d decompose and reconstruct seem to be working well
2016-08-14 16:29:27 -06:00
Samuel Li
53cfd74d01
Rectangle Copy test succeed
2016-08-14 00:28:29 -06:00
Samuel Li
9e49cde593
2D decomposition work in progress
2016-08-11 18:49:48 -06:00
Samuel Li
be658d70a8
IDWT2D seems to be working
2016-08-11 15:24:48 -06:00
Robert Maynard
04761975be
Refactor FieldHistogram so class is not templated.
...
Move the template parameters to the Run method.
2016-08-11 14:46:45 -04:00
Samuel Li
d41d102072
DWT2D works
2016-08-10 16:28:59 -07:00
Samuel Li
710d59d33a
Change on class signatures of Filter and Base
2016-08-10 17:24:22 -06:00
Samuel Li
e7b288dba1
eliminates explicit calling VTKM_DEFAULT_DEVICE_ADAPTER_TAG
2016-08-10 14:33:14 -06:00
Samuel Li
f669f44ca1
new branch on 2D wavelets
2016-08-09 17:32:13 -06:00
Samuel Li
f3d7aeced1
use std::vector<vtkm::Id> instead of raw arrays for bookkeeping
2016-08-08 17:59:15 -06:00
Samuel Li
c72b135608
Uncomment other build objects
2016-08-08 12:21:08 -04:00
Samuel Li
34ce28576d
change unittest for review
2016-08-06 12:34:03 -07:00
Samuel Li
3d29e504cc
fixed a memory bug, now I believe it's ready for merge
2016-08-06 00:08:29 -07:00
Robert Maynard
08a33675f6
Prefer vtkm::cont::CastAndCall function over the member method.
2016-08-05 15:29:39 -04:00
Robert Maynard
31138293f9
Pass Fields and CoordinateSystems to Dispatcher::Invoke
...
Now that we can handle scheduling Fields and CoordinateSystems, do
2016-08-05 15:26:17 -04:00
T.J. Corona
927387b1a2
Add vtkm::cont::Field,CoordinateSystem as Invokable parameters.
...
Originally, only array handles could be passed to the dispatcher.
This update extends the list to Field and CoordinateSystem as
dynamic objects.
2016-08-03 15:46:30 -04:00
Samuel Li
448e9be5a8
small tweaks, ready for pull request
2016-07-29 14:16:10 -07:00
Samuel Li
dd1ef8e47d
include all objects in the cmake file
2016-07-29 13:19:28 -07:00
Samuel Li
3cc6385426
remove unnecessary files
2016-07-29 13:05:53 -07:00
Samuel Li
ec46fd58a0
fixed bugs, ready for Chris to review
2016-07-29 11:27:37 -07:00
Samuel Li
7f9fd0d66d
finish all the statics
2016-07-28 21:07:05 -06:00
Samuel Li
fa1198bb85
summation algorithm function is ready
2016-07-28 19:12:57 -06:00
Samuel Li
1a7bae0335
with a hack on ArrayHandlePermutation.h, wavelet works on CUDA device as well
2016-07-28 17:55:38 -06:00
Samuel Li
951d9fce2c
move device copy to a function of itself; other small changes
2016-07-28 14:55:10 -07:00
Robert Maynard
73440ccb4e
vtkm::filter::MarchingCubes now uses vtkm::worklet
...
Instead of having two separate implementations of MarchingCubes, we have
a single implementation in worklet, and filter uses that.
That does mean that the version in worklet has been updated to support
merging of duplicate point coordinates.
2016-07-28 08:52:34 -04:00
Samuel Li
7dc59d80d7
trying to figure out how to use deviceAdapterAlgorithms
2016-07-27 16:23:26 -07:00