summaryrefslogtreecommitdiffstats
path: root/cuda
Commit message (Collapse)AuthorAgeFilesLines
* Clean up unused variablesWillem Jan Palenstijn2016-10-131-29/+0
|
* Increase max angle count for Parker WeightingWillem Jan Palenstijn2016-10-131-1/+1
|
* Merge branch 'master' into FDKWillem Jan Palenstijn2016-10-061-39/+31
|\
| * Compute FBP filter in spatial domainWillem Jan Palenstijn2016-06-241-7/+39
| |
* | Merge branch 'master' into FDKWillem Jan Palenstijn2016-04-221-18/+13
|\|
* | Use CompositeGeometryManager for FDKWillem Jan Palenstijn2016-04-188-245/+176
| | | | | | | | | | Also fix a number of scaling/weighting issues in FDK, and switch to standard cone_bp with FDKWeighting for the BP step.
* | Add FDKWeighting option to standard cone_bpWillem Jan Palenstijn2016-04-182-4/+12
| |
* | Merge branch 'master' into anisoWillem Jan Palenstijn2016-04-1810-3/+68
|\|
| * Add relaxation parameters to SIRT3DWillem Jan Palenstijn2016-03-234-1/+27
| |
| * Add relaxation parameters to SIRT, SARTWillem Jan Palenstijn2016-03-234-2/+21
| |
| * Query max texture size instead of hardcoding itWillem Jan Palenstijn2016-02-162-0/+20
| |
* | Adapt 3D BP scaling to voxel sizeWillem Jan Palenstijn2016-04-182-4/+8
| |
* | Move detector-independent vars out of loopWillem Jan Palenstijn2016-04-181-14/+12
| |
* | Add voxel-size scaling to par3d_fpWillem Jan Palenstijn2016-04-181-16/+92
| |
* | Refactor voxel size scaling CUDA kernel, and special-case cubesWillem Jan Palenstijn2016-04-181-26/+62
| |
* | Process non-cubic-voxel astra geometriesWillem Jan Palenstijn2016-02-101-4/+11
| |
* | Add cone_fp kernel support for anisotropic voxelsWillem Jan Palenstijn2016-02-102-10/+31
| |
* | Refactor CUDA projector params into structWillem Jan Palenstijn2016-02-1017-190/+185
|/
* Add multi-GPU support to CompositeGeometryManagerWillem Jan Palenstijn2016-01-191-0/+2
|
* Reduce FP3D CUDA kernel runtimeWillem Jan Palenstijn2015-12-222-2/+2
| | | | | This reduces the chance of the Windows display driver watchdog triggering, and doesn't seem to hurt performance.
* Add CompositeGeometryManagerWillem Jan Palenstijn2015-12-044-82/+378
| | | | | This handles FP and BP operations on multiple data objects at once, splitting them to fit in GPU memory where necessary.
* Fix assertWillem Jan Palenstijn2015-07-161-1/+1
|
* Merge branch 'master' into volgeom3dWillem Jan Palenstijn2015-04-158-44/+47
|\
| * Fixed a few CUDA 2D DART bugs.Valerii Sokolov2015-04-091-6/+7
| | | | | | | | | | * Mixed width and height led to incorrect work on rectangular images. * Incorrect weight calculation in `devDartSmoothingRadius` (#47).
| * Initialize variables to avoid warningWillem Jan Palenstijn2015-03-303-3/+3
| | | | | | | | | | These variables are never used when uninitialized, but Visual Studio complains about them.
| * Fix indentationWillem Jan Palenstijn2015-03-201-11/+11
| |
| * Adds ASTRA_*** defines for easier logging, and changes internal calls to ↵Daniel M. Pelt2015-03-134-21/+21
| | | | | | | | these defines
| * Use new logging API internally instead of printf / iostreamDaniel M. Pelt2015-03-104-24/+31
| |
| * Remove old Logging code (only used in fft.cu)Daniel M. Pelt2015-03-092-9/+4
| |
* | Add outputScale argument to 2D CUDA BPWillem Jan Palenstijn2015-03-1212-63/+76
| |
* | Adapt standalone test programs to outputscaleWillem Jan Palenstijn2015-03-122-2/+2
| |
* | Scale CUDA 3D FP/BP output with volume pixel sizeWillem Jan Palenstijn2015-03-125-37/+36
| |
* | Add outputScale argument to 3D CUDA BPWillem Jan Palenstijn2015-03-129-38/+54
| |
* | Add flexible volgeom3d support to astraCudaFPWillem Jan Palenstijn2015-03-122-217/+36
| |
* | Add flexible volgeom3d support to astraCudaBPWillem Jan Palenstijn2015-03-112-202/+21
| |
* | Add flexible volgeom3d support to astraCudaBP_SIRTWeightedWillem Jan Palenstijn2015-03-112-80/+38
| |
* | Add another utility functionWillem Jan Palenstijn2015-03-111-46/+53
| |
* | Let astraCudaFDK use utility functionsWillem Jan Palenstijn2015-03-112-33/+16
| |
* | Adjust interface slightlyWillem Jan Palenstijn2015-03-111-11/+12
| |
* | Allow non-centered volume geometry in SIRT3D and CGLS3DWillem Jan Palenstijn2015-03-092-286/+88
| |
* | Add 3d geometry normalization functionsWillem Jan Palenstijn2015-03-062-0/+166
|/
* Untangle cuda and astra includesWillem Jan Palenstijn2015-03-023-123/+4
|
* Greatly speed up CUDA par3d_bpWillem Jan Palenstijn2015-01-291-53/+48
| | | | These are the changes to cone_bp applied to par3d_bp.
* Greatly speed up CUDA cone_bpWillem Jan Palenstijn2015-01-291-99/+96
| | | | | This improves speed by a factor of 10. Tested on Kepler. Joint with Jeroen Bédorf.
* Update copyright to -2015Willem Jan Palenstijn2015-01-2655-110/+110
|
* Update copyright headersWillem Jan Palenstijn2015-01-1955-275/+275
|
* Remove unused filesWillem Jan Palenstijn2014-12-042-177/+0
|
* Move BP coordinate transformation to utility functionWillem Jan Palenstijn2014-12-034-0/+89
|
* Start supporting flexible 2D volume geometry in Cuda FBP tooWillem Jan Palenstijn2014-10-062-25/+14
|
* Add CUDA SIRT::doSlabCorrections() functionWillem Jan Palenstijn2014-10-024-0/+59
| | | | | This function optionally compensates for effectively infinitely large slab-like objects of finite thickness 1.