Kernel reimplementation and bug fixes (#1)

* Init refactor

* Add correlator interface

* Update options file

* Implement alloc and reset functions

* Implement constructor and distructors

* Add basic main

* Add function call to kernel

* Partially Implement correlation kernel

* Improve help message

* Improve help message

* Continue kernel implementation

* Continue kernel re-implementation

* Continue kernel re-implementation

* Finish kernel implementation (hopefully)

* Add transfer function

* Remove unused macros

* Add debug build

* Update debug build

* Add utils

* Update options

* Update example main

* Fix bug in memory allocation

* Fix memory alignment warning

* Partial fix correlate kernel

* Reimplement kernel...

* Implement kernel

* Finish kernel implementation

* Update debug build

* Update get function and generate taus

* Fix bug in autocorrelation kernel

* Add timing to main

* Update main

* Update options

* Naming consistency

* Update run_all.sh script

* Update run_all.sh script

* Remove num_Sensors_per_block as parameter

* Fix kernel implementation

* Update main

* Fix kernel implementation

* Update output (temporary to checks results)

* Update tau generation

* Update output results (exec time are temporary)

* Fix kernel (maybe)

* Update results

* Update outputs

* Update run all script

* Memory optimization

* Update results

* Remove useless barriers

* Update results

* Improve insert_until_bin kernel

* Change data types for less memory usage and faster controls

* Update results

* Improve debug info

* Update stuff

* Remove warning print

* Update results

* Fix bugs

* Fix memory misalignemt

* Update outputs

* Update time results

* Update timing results for 30000 (64 bit needed)

* Update run configuration

* Add comments

Author: feDann

.gitignore

@@ -3,4 +3,6 @@ out/
 *.vcxproj.*
 *.txt
 main
-documentation/Doxygen/
+documentation/Doxygen/
+bin/
+.vscode/

Makefile

@@ -1,17 +1,21 @@
-NVCC=nvcc
-CXX=g++
+CXX=nvcc
 
-CXXFLAGS=-Ofast -std=c++11 -D_MWAITXINTRIN_H_INCLUDED -D_FORCE_INLINES
-LIBS=-lcudart
-FLAGS=-Xcompiler -fopenmp -D_MWAITXINTRIN_H_INCLUDED -D_FORCE_INLINES
+CUDAFLAGS=--ptxas-options=-v -m64 -arch compute_61 -code sm_61 -Xptxas -dlcm=ca -Xcompiler -D_FORCE_INLINES -lineinfo --expt-relaxed-constexpr
 
-CUDAFLAGS=--ptxas-options=-v -O4 -m64 -arch compute_61 -code sm_61 -Xptxas -dlcm=ca -Xcompiler -D_FORCE_INLINES -lineinfo
+BIN_FOLDER=bin
+SRC_FOLDER=src
 
-all: main
+FILES=${SRC_FOLDER}/main.cu ${SRC_FOLDER}/correlator.cu
+
+all: release
 
 clean:
-	rm -f *.o main out_data.txt timer_out.txt
+	rm  -rf $(BIN_FOLDER)
 
-main: src/CudaInput.h src/InputVector.h src/BinGroupsMultiSensorMemory.h  src/DataFile.h  src/Timer.h src/Main.cu src/SensorsDataPacket.h src/ResultArray.h src/options.hpp src/utils.hpp src/Definitions.h
-	$(NVCC) $(CUDAFLAGS) src/Main.cu -o main
+release:
+	mkdir -p $(BIN_FOLDER)
+	$(CXX) $(FILES) $(CUDAFLAGS) -O3 -o $(BIN_FOLDER)/main
 
+debug:
+	mkdir -p $(BIN_FOLDER)
+	$(CXX) $(FILES) $(CUDAFLAGS) -D_DEBUG_BUILD -g -G -o $(BIN_FOLDER)/main 

documentation/AlgorithmPresentation.pptx

@@ -1,30 +1,30 @@
-3.0204e-05
-2.7346e-05
-2.7161e-05
-2.6487e-05
-2.6146e-05
-2.652e-05
-2.7012e-05
-4.4106e-05
-2.6444e-05
-2.6492e-05
-2.679e-05
-2.5983e-05
-2.6658e-05
-4.1025e-05
-2.6642e-05
-2.6806e-05
-2.5854e-05
-3.4931e-05
-2.6795e-05
-2.9797e-05
-3.438e-05
-2.6897e-05
-3.7489e-05
-3.9591e-05
-4.1523e-05
-2.6488e-05
-2.676e-05
-3.6691e-05
-2.6617e-05
-2.6519e-05
+3.2256e-05
+4.0013e-05
+4.7933e-05
+3.7918e-05
+6.3048e-05
+4.7005e-05
+3.1938e-05
+3.1058e-05
+4.5578e-05
+2.4995e-05
+3.7842e-05
+3.7918e-05
+3.8048e-05
+2.9517e-05
+3.8357e-05
+3.7484e-05
+3.0092e-05
+3.7669e-05
+3.753e-05
+3.7721e-05
+3.7534e-05
+3.8252e-05
+3.942e-05
+3.2329e-05
+3.1416e-05
+3.7708e-05
+3.769e-05
+3.6285e-05
+3.6966e-05
+4.6296e-05