https://github.com/wavestoweather/PCViewer.git

sort by:
Revision Author Date Message Commit Date
58169f6 Trying out non atomic increment. 08 June 2022, 11:13:43 UTC
67afe0e A lot of bugfixes... 07 June 2022, 17:19:04 UTC
2c432b3 Axis scaling now works, added activation to the cpu implementations, recalculation now also works. 04 June 2022, 08:31:19 UTC
5d802cd Huge improvement for hierarchy creation, added index activation and axis scaling. 03 June 2022, 17:40:01 UTC
f7c44cb Found better compression reasons in the previos id files. Compression time is very slow: test for normla vector list with following insertion into roaring bitmaps. 02 June 2022, 17:44:36 UTC
d94d9ce Added cpu and gpu code for rle decoding. Still has to be fully implemented and integrated. 01 June 2022, 17:04:54 UTC
c1f3664 Added first cpu side things for Huffman decoding. Still missing pipeline setup and other things. 31 May 2022, 17:21:38 UTC
ab4e336 Added huffman decode transpose shader 31 May 2022, 16:42:04 UTC
bc14a55 Added huffman decoding gpu code 31 May 2022, 14:07:10 UTC
2edca73 Huffman table cpu side mostly implemented. GPU side missing. 27 May 2022, 10:49:28 UTC
0088689 Started to implement vkCompress to be able to decompress the datablocks on the fly. 27 May 2022, 08:48:02 UTC
840b9b3 lineCountAll functions now without problems, bug when wanting to close the application fixed. For 6.6 million datapoints coutn for all attributes takes 340ms. 26 May 2022, 17:44:26 UTC
00c0454 Fixed bug in data.hpp, added new subsampling strategy for laoding (random sampling). 26 May 2022, 13:20:45 UTC
4993df1 GPU pipelines are connected and are working. All once compute pipeline currently under development. 25 May 2022, 17:04:32 UTC
4d0b390 Added dropdown menu to change counting strategy. 24 May 2022, 20:03:22 UTC
a5fac25 Test dataset with 10 mio datapoints counted with cpu approach now works including rendering. 24 May 2022, 19:45:52 UTC
3a50a04 Compilable adoption of indbinmanager to PCViewer.cpp. Started creation of the compression method to produce the correct format. 23 May 2022, 19:56:09 UTC
d8fbc13 Added safety checks at pipelines plus freeing of command buffers. Project compilable again. 23 May 2022, 13:35:33 UTC
d871212 Added renderer ot the IndBinManager. Next: add IndBinManager to drawlist. 23 May 2022, 13:28:35 UTC
8c0d365 Added gpu pairwise counting pipelines to the bin manager. Render counter not yet completed (missing float to uint conversion pipeline). 23 May 2022, 12:27:27 UTC
9859bd6 Cpu counting adopted hooked up to IndBinManager. 23 May 2022, 08:31:15 UTC
29824d7 Added automatic half conversion, cpu counting added. Other connections are still missing. 19 May 2022, 19:46:14 UTC
d73a04a Implemented loading and conversion of 1d indexlists. Compilable, not yet debugged. Next: Data loader for compressed and uncompressed formats, GpuDecompression. 18 May 2022, 17:21:49 UTC
12bd11b Corrected benchmarktests for lineCounter for proper non partitioned timings. 18 May 2022, 14:10:06 UTC
f0bf245 Further implemented counting. 17 May 2022, 18:23:22 UTC
83c3bcf Started implementation of general load manager for 2d binning and rendering. 12 May 2022, 18:12:54 UTC
4db50dd Added renderer for bin counts. Compilable but not yet fully tested with true data. 09 May 2022, 18:06:04 UTC
f048783 Fixed compile warnigns. 28 April 2022, 15:18:38 UTC
ffec189 Added extra test file for faster compilation when changing tests, tried faster compute shader, added roar bitmap tests for intersection calculation, added range class for more beatiful irange iteration, commented out unnecessary things in hierarchy bin manager and tested real world performance of roaring bitmaps on over half a billion datapoints. 27 April 2022, 18:37:31 UTC
672193b Test code for roaring bitmaps: Avg compression ratio for real world data was at 1:~32 26 April 2022, 13:30:44 UTC
17c957f Added roaring bitmaps for byte size tests. 13 April 2022, 13:50:58 UTC
a74e7cd Code futher implemented for timings 12 April 2022, 16:20:32 UTC
bc49d9c Added a cpu counting to the pcviewer for performance comparison. 08 April 2022, 09:24:10 UTC
3b62f44 Added rendering counting pipeline. For 2^30 datapoints 800 ms are needed for computation. 07 April 2022, 18:53:17 UTC
b3c3a01 2^30 points test with good spread -> 1.4s for counting between two attributes. 07 April 2022, 15:54:29 UTC
0c8adc9 First implementation using compute shader for line counting done. Does not work for very large data sizes. 07 April 2022, 15:18:06 UTC
ce5f744 Started implementation of counting pipeline. 07 April 2022, 08:24:52 UTC
f05b476 Better stopwatch, better combination calculation. 01 April 2022, 14:00:57 UTC
058b4a3 Working implementation for cluster counting. Will get a runtime improvement. 01 April 2022, 09:02:20 UTC
71c632d Fixed build for ubuntu. 28 March 2022, 08:20:11 UTC
7f41046 Further implemented counting active lines. 27 March 2022, 17:27:25 UTC
414a80f Started to implement loading and conversion to drawable line instances. 22 March 2022, 18:57:00 UTC
e12fb1b Started implementation of bin loading and conversion manager. 22 March 2022, 14:07:55 UTC
b32ea8a 1d cluster indices implementation done. 22 March 2022, 13:38:34 UTC
7357030 Further implemented rendering. Data conversion to correct format done, now rendering with indexbuffercreation can be implemented. Only testing is missing then. 17 March 2022, 18:54:05 UTC
7164438 Initial commit for rendering pipeline. Runs without errors. 17 March 2022, 10:07:29 UTC
93d98cb Loading done and debugged. 16 March 2022, 20:09:42 UTC
b7bab3d First implementation for loading new hierarchys done. Still has to be debugged. 16 March 2022, 09:33:33 UTC
c7baadd Started implementing HierarchyLoadManager to load the subdimension clusters. 15 March 2022, 20:11:44 UTC
7ef5216 Disabled stopwatches 14 March 2022, 18:56:57 UTC
fe36de8 Tried to do multithreading speedup. 14 March 2022, 18:37:48 UTC
3f3568b Starte implementation of nd clustering. 11 March 2022, 19:22:24 UTC
159c74b Inbetween commit. 10 March 2022, 18:36:11 UTC
4175958 Started implementation of ndHierarchy creation. Permutation creation for n dimensions done. 10 March 2022, 13:37:44 UTC
b7c2ee8 Finished implementation and timing of hashLeaderNode. Hash vector leader node is superior to hashLeaderNode. Timings for hash vector leader node was commented out for possible future tests. 08 March 2022, 09:51:39 UTC
5907d75 Added hash map accelerated hierarchy creation. Timing watches are still included in this commit, should be removed. Hash map acceleration is faster. Implementation for hash map accelerated reference pointer begun, still has to be finished and timed. 04 March 2022, 17:17:10 UTC
bc45f2c Fixed bundle reloading. 03 March 2022, 12:15:53 UTC
257929e Implemented non compressing pipeline for comparison. 03 March 2022, 09:40:22 UTC
52ef13c Hierarchy creation and compression fixed, hierarchy loading adapted. 02 March 2022, 20:44:54 UTC
67a5789 Hierarchy import manager adopted to be able to load the new hierarchy files. 25 February 2022, 18:56:30 UTC
8cff003 Started to implement decompression of bundled data. 25 February 2022, 10:09:59 UTC
b79bbed Bundled compression done. 25 February 2022, 09:16:28 UTC
20ad471 Added the ability to get all needed cache nodes out of the hierarchy in a sinlge pass including caching these file in a single pass. 24 February 2022, 20:25:33 UTC
844b1a7 Started to add more advanced caching methods with an cache manager interface to be able to exchange the caching backend. Also added a new leaders clustering approach, namely vector leader node which has only vectors to hold leaders information for more efficiency. In HierarchyCreation the new methods for the new caching manager to compress the temporary files has been begun to be implemented. 23 February 2022, 19:23:29 UTC
971faac Compilable under windows, still has some memleaks. 17 February 2022, 13:16:55 UTC
6e16e54 Caching bug solved. 17 February 2022, 08:29:20 UTC
a7d064d A lot of bugfixes for compression workbench as well as runtime improvements. Implementation fo a HashNode which tries to increase the runtime, is however slower as the leaders node. Further the adoption of the hierarchy import manager to account for brushing was attempted, some error still exists. 16 February 2022, 18:00:05 UTC
87cc287 Fixed small bug in hierarchy creation, implemented adoption to drawlist to be able to use the hierarchy manager reloading to properly show the dataset. 14 February 2022, 13:42:12 UTC
dd4b3de Changed drawlist creation to be able to handle hierarchical files. Hierarchy building seems to have some problem(combination of compressed files). 13 February 2022, 18:07:40 UTC
7ffdb9f Compilable version. Continuing to adopt create drawlist. 13 February 2022, 07:45:21 UTC
542811b Opening of hierarchy in the PCViewer was added. Creation and drawing of hierarchical drawlists with automatic reloading not yet implemented. 12 February 2022, 19:15:41 UTC
e1371e1 Started adoptions for main pcp rendering to be able to also include changin drawlist data. 11 February 2022, 18:22:07 UTC
0c849c1 bugfix for inBrush function. 11 February 2022, 18:01:49 UTC
12119bc First iplementation for the Hierarchy import manger to automatically determine the hierarchy levels done. Compilable, but not yet tested. 11 February 2022, 18:01:27 UTC
2038d61 Rtree in leaders clustering is now static for up to 30 dimensions, above rtree has nodes with dynamic size dimension arrays. Not yet fully tested and debugged. 09 February 2022, 20:06:01 UTC
4eb5e5a Added rtree util with the variant. 09 February 2022, 18:15:29 UTC
7462a19 Changing rtree to a rtree variant which adopts to the amount of dimensions (uses a static version of rtree which is faster). 09 February 2022, 18:15:04 UTC
0e567c8 Compile fixes under ubuntu. 08 February 2022, 13:22:29 UTC
91fc68a Added documentation for hierarchical file layouts and compression methods. 07 February 2022, 17:31:44 UTC
d9efd8b Started import manager implementation with automatic adoption to brushes. 04 February 2022, 17:54:54 UTC
7ec073a Implemented merging of temporary clusters with following compression via dwt, quantization, run length encoding and last huffman encoding. Correct writeout to files as well as correct import of compressed file. The compression also automatically converts rwo major points to column major to be able to faster convert to data object and to have more efficient compression. 03 February 2022, 18:31:13 UTC
95c6a8a Dwt as well as quantization are working, cpu compression still has some errors. Main contains debug code. 02 February 2022, 20:30:31 UTC
ce6f95e Added abstract class for clustering nodes. 02 February 2022, 08:10:17 UTC
2a3a6a3 Added cudaCompress for cpu to encode the cache files. Rework was done so that only the cpu version of cuda compress is available and is compilable when all files are inthe same file. 29 January 2022, 18:48:16 UTC
a3456af Deadlock fix, more info in the gui, file writeout fixed, multhtread implementation nearly completly done (Missing proper canelation fo hierarchy building). 29 January 2022, 18:10:49 UTC
241a549 Added caching if ram is being filled up. Hierarchy creation is now able to run multithreaded. Thread amount is settable, insertions are lock free and secure. Writing out the chaceh files has still some errors. 28 January 2022, 18:18:58 UTC
ff3554d bugfixes for hierarchy creation. 27 January 2022, 19:20:49 UTC
ee6e6a8 Implemented Hierarchy Node to be able to recursiveley add datapoints. First take on tmp hirarchy creation implemented using rtrees as acceleration structures. Current version compilable without errors and warnings. 27 January 2022, 10:16:15 UTC
e2b177c Compression workbench is now usable from within the pcviewer application. Data anlysis is moved to a separate thread. 25 January 2022, 15:10:02 UTC
6388e04 Fixed small bugs in loader, implemented new compression workbench gui to do data loading in the PCViewer. 25 January 2022, 13:56:30 UTC
771e1c3 Started implementatio of Hirarchy building. NetCdf loader nearly done. 21 January 2022, 19:20:52 UTC
c685d1b Tested framebuffer addition with a single component float framebuffer-> same render times as normal rendering if opacity = 1/255, else slower(but constantly slow). 18 January 2022, 17:14:31 UTC
1c5ddb0 Working implementation for alternative parallel coordinates rendering via atomic add on image with a resolve pass. Unfortunately around 10 times slower. Improvement: use float buffer to be able to use output merger again. 18 January 2022, 15:00:43 UTC
c16a68e Compilable and runnable version. Does not yet work as the image has to be added to the fragment shader stage via descriptor set to be able to atomic add into the image. 17 January 2022, 18:16:52 UTC
cf6ff49 First done implementation. Not yet checked for compilation and bugs. 17 January 2022, 17:04:31 UTC
16b4e66 Inbetween commit for efficient rendering pipeline. 17 January 2022, 14:01:15 UTC
31c9595 Further implemented dataset gpu sorting. 07 January 2022, 16:50:08 UTC
255f3f8 Added data access to dataset index sorting. Sorting still has to be adopted to use the dataset. 05 January 2022, 10:01:14 UTC
115e137 Hybrid radix sort improvement by better loading strategy. 03 January 2022, 15:06:05 UTC
861e621 Gpu radix sorter improvement, started to add things for dataset sorting. 29 December 2021, 09:52:18 UTC
back to top