aboutsummaryrefslogtreecommitdiff
path: root/src/diffraction-gpu.c
AgeCommit message (Collapse)Author
2012-02-22Remove all remaining PDB rendering stuffThomas White
2012-02-22Add tests/gpu_sim_checkThomas White
2012-02-22setup_gpu() does not need an imageThomas White
2012-02-22Use LUTs for CPU simulation as wellThomas White
2012-02-22Check errors properlyThomas White
2012-02-22Make the GPU simulation workThomas White
2012-02-22Tidy up, fix definition of get_q() in CL kernelThomas White
2012-02-22Fix GPU code for new geometry, and tidy up some detector stuff (needs debugging)Thomas White
2012-02-22Use symmetry when simulating on the GPUThomas White
2012-02-22Use symmetry when simulating (on the CPU only)Thomas White
2012-02-22Handle multiple GPUsThomas White
2012-02-22Fix previous commitThomas White
2012-02-22Zap quatsThomas White
The "orientation" quaternion should not rotate the Ewald sphere, but the crystal. This cleans things up a little, removing "orientation" from "struct image".
2012-02-22pattern_sim: Fix memory leaksThomas White
2012-02-22New flat value when no structure factors are availableThomas White
2012-02-22Fix wavelengthThomas White
2012-02-22Update definition of bandwidth to match documentationThomas White
2012-02-22Add "bandwidth" to beam parameters fileThomas White
Also, fix previous commit.
2012-02-22Reduce the scope of "count"Thomas White
Lists of counts had pervaded every corner of CrystFEL, being used as markers for the presence of reflections. Now we have a better way of doing this, the ReflItemList, and few parts of the suite apart from process_hkl have any business knowing how many observations were made of a particular reflection.
2012-02-22pattern_sim: Fix up GPU diffraction for new detector geometryThomas White
2010-04-20Use new access method for OpenCL vectorsThomas White
2010-04-01indexamajig: Introduce multithreadingThomas White
2010-03-27Make it obvious, when using GPU simulation, if a required reflection is ↵Thomas White
unavailable
2010-03-26Switch bandwidth off for nowThomas White
2010-03-26Don't try to render PDBs, part II: remove "molecule" from "struct image".Thomas White
2010-03-26Don't try to render PDBs, part IThomas White
2010-03-10Remove debugging messagesThomas White
2010-03-10Tidy upThomas White
2010-03-10Allow multiple sinc LUTsThomas White
2010-03-02Clean up sinc LUTsThomas White
2010-03-02Take modulus when calculating sinc LUTThomas White
2010-03-01Use a lookup table for sinc values in GPU calculationThomas White
2010-02-27More tests for odd situationsThomas White
2010-02-27Fix type conversions to avoid trouble with overflows etc.Thomas White
2010-02-26Move water calculation to diffraction.c, and work out the consequencesThomas White
2010-02-24Report maximum work group size and if it has been exceededThomas White
2010-02-23Tidy up console outputThomas White
2010-02-23Mac OS X interoperabilityThomas White
2010-02-22Don't try to calculate with a null GPU contextThomas White
2010-02-22Report when setting up GPUThomas White
2010-02-20Move OpenCL utility stuff to a separate moduleThomas White
2010-02-20Nicer branchingThomas White
2010-02-20Use the right value for --no-sfacThomas White
2010-02-19Move critical array size definitions to share between CPU and GPU versionsThomas White
2010-02-19Add bandwidth to GPU calculationThomas White
Also: alter CPU version to be cleaner and give exactly the same results at GPU, and fix an indexing bug
2010-02-19Fix various exciting (and otherwise) memory leaksThomas White
2010-02-19Simplify sampling a bit - preventing some indexing issuesThomas White
2010-02-19Add finite sampling for better simulation accuracyThomas White
2010-02-19Synchronise properlyThomas White
2010-02-19Split OpenCL initialisation into separate routing to avoid re-compiling all ↵Thomas White
the time