diff --git a/makefile~ b/makefile~
deleted file mode 100644
index 48fb814..0000000
--- a/makefile~
+++ /dev/null
@@ -1,67 +0,0 @@
-#CXX=g++ -g -Wall  #for debug mode
-CXX=g++ -O3 -Wall
-### -O2 -O5 -Os
-#g++ `pkg-config --cflags itpp` -o hello.out hello.cpp `pkg-config --libs itpp`
-
-
-
-START=`pkg-config --cflags itpp`
-END=`pkg-config --libs itpp`
-#files=mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h makefile
-#command=$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-
-
-#update:
-INC_DIR=~/working/weilei_lib
-files=$(INC_DIR)/mm_read.c $(INC_DIR)/mm_read.h $(INC_DIR)/mmio.c $(INC_DIR)/mmio.h $(INC_DIR)/mm_write.c $(INC_DIR)/mm_write.h $(INC_DIR)/lib.cpp $(INC_DIR)/lib.h $(INC_DIR)/dist.c $(INC_DIR)/dist.h $(INC_DIR)/concatenation_lib.c $(INC_DIR)/concatenation_lib.h $(INC_DIR)/my_lib.h makefile
-#command=$(CXX) -I../../weilei_lib $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(word 10, $^) $(word 12, $^) $(END)
-
-command=$(CXX) -I $(INC_DIR) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(word 10, $^) $(word 12, $^) $(END) ; echo finish making
-
-
-
-all: parserTest.out ldpcTest.out test.out alist_test.out
-###include all headfiles into my_lib.h
-#pattern_observer.out:pattern_observer.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h makefile
-partial_sum.out:partial_sum.c $(files)
-	$(command)
-pattern_observer.out:pattern_observer.c $(files)
-	$(command)
-bp_decoding4.out:bp_decoding4.c $(files)
-	$(command)
-my_bp1.out:my_bp1.c $(files)
-	$(command)
-#bp_decoding3.out:bp_decoding3.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h makefile
-bp_decoding3.out:bp_decoding3.c $(files)
-	$(command)
-data_collect_perm_check.out:data_collect_perm_check.c $(files)
-	$(command)
-convert_data.out:convert_data.c $(files)
-	$(command)
-test.out:test.c $(files)
-	$(command)
-	./test.out
-#	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(END)
-error_analysis.out:error_analysis.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(END)
-rand_decode_perm_check.out:rand_decode_perm_check.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-rand_decode_perm.out:rand_decode_perm.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-rand_decode3.out:rand_decode3.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-rand_decode.out:rand_decode.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-gauge_to_stabilizer.out:gauge_to_stabilizer.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(END)
-decoding.out:decoding.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h my_lib.h
-	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(END)
-#test.out:test.c mm_read.c mm_read.h mmio.c mmio.h mm_write.c mm_write.h lib.cpp lib.h my_lib.h
-#	$(CXX) $(START) -o $@ $< $(word 2,$^) $(word 4, $^) $(word 6, $^) $(word 8, $^) $(END)
-code_generator.out:code_generator.c mmio.c mmio.h mm_write.c mm_write.h
-	$(CXX) $(START) -o $@ $< $(word 2, $^) $(word 4, $^) $(END)
-clean:
-	echo this command clean all temperary files in this folder.
-	rm  *~
-	rm \#*
-
diff --git a/my_bp1.c b/my_bp1.c
index eb0a62c..2cf26c5 100644
--- a/my_bp1.c
+++ b/my_bp1.c
@@ -1,15 +1,5 @@
-//updated on Dec 9.
-//using optimizations (compared with bp_decoding3.c)
-// bits_out = llr_output < bound,
-// repeat decoding
-//the input become syndrome, not the error vector. (this should not be the difference or produce any differenc ein the result. If I have time, I will implement it in bp_decoding3.c
-//exit at 200 ant repeat 5 times
-
-//In order to change to z error only, there is no change needed for this code. Just use the new generating matrix. 2/10/2018
-//Compared to bp_decoding2.c, in this file,I save all data for errors that converge or not converge in bp decoding.
-//In this file, for toric code, we use stabilizer S=(S_x,S_z).
-//errors are in the format e=(e_z,e_x). (bistin, recbits, bitsout) //not checked excatly, based on the assumption that the BP decoding in itpp doesnot know the difference between X and Z errors and operators.
-//hello world
+//Weilei  March 26, 2020. yse c++ async to manage multi threads
+// copied from bp_decoding4.c
 
 #include <itpp/itbase.h>
 #include <itpp/itcomm.h>
@@ -30,8 +20,6 @@ int main(int argc, char **argv){
   char * filename_result=argv[3];//prefix for the file
   double p=atof(argv[4]);
   p=p/100000.0;//previous use 1000 division. Now use 100,000 division cause the thershold for toric codes seems to be around 0.1%.
-  //  cout<<"prob p ="<<p<<endl;
-  //cout<<"input file -->"<<filename_G<<endl;
 
   //parameter setup
   int cycles=3000;//10000;//number of cycles: fro toric code, 10000 give reletively clear result
@@ -51,9 +39,6 @@ int main(int argc, char **argv){
   //    LDPC_Code C=MM_to_LDPC_Code(filename_G);  //convert GF2mat saved in .mm file to LDPC_Code
   //LDPC_Code C = get_test_LDPC_Code();
   
-  //LDPC_Code C(filename_G);//load code if saved in .it file with LDPC_Code format
-  //vec pvals = "0.01:0.01:0.1"; p.get(pvals,"pvals");// start:increment:end=inclusive
-
   //bp decoding set up
   C.set_exit_conditions(exit_at_iteration,true,true);//high perperformance. This number of iteration would not affect small errors which converges and stop very fast. Only affect those doesn't converge soon or not converge at all.
   //  C.set_exit_conditions(50,true,true);  // 50 iterations,check syndrome always
@@ -234,26 +219,7 @@ cout<<"rec_bits"<<endl;
   cout<<", Converge rate ="<<rate_converge<<endl;
   //  save_result(p,rate_converge,filename_result_p);//no need to save this. can get it by counting the size of the matrix when doing gnuplot
   
-  //save errors to files
-  char filename_result_ic[255];
-  sprintf( filename_result_ic,"%s%.5f_input_converge",filename_result,p);//append p to the file name
-  //cout<<E_input_converge.rows()<<endl;
-  GF2mat_to_MM(E_input_converge,filename_result_ic);  
-    
-  char filename_result_in[255];
-  sprintf( filename_result_in,"%s%.5f_input_nonconverge",filename_result,p);//append p to the file name
-  GF2mat_to_MM(E_input_nonconverge,filename_result_in);  
 
-  char filename_result_oc[255];
-  sprintf( filename_result_oc,"%s%.5f_output_converge",filename_result,p);//append p to the file name
-  GF2mat_to_MM(E_output_converge,filename_result_oc);  
-
-  char filename_result_on[255];
-  sprintf( filename_result_on,"%s%.5f_output_nonconverge",filename_result,p);//append p to the file name
-  GF2mat_to_MM(E_output_nonconverge,filename_result_on);  
-  
-  //  cout<<E_input_nonconverge<<endl;
-  // cout<<E_input_converge<<endl;
   timer.toc_print();
   return 0;
 }
diff --git a/run_save_bp4.sh b/run_save_bp4.sh
new file mode 100755
index 0000000..33a1d33
--- /dev/null
+++ b/run_save_bp4.sh
@@ -0,0 +1,51 @@
+#runs bp_decoding3.out and save all the input and output errors
+
+#try bp decoding for toric code with different size, 5,9,25,35
+#added size 7,11,13
+#p=0.1% to 5%
+#100,000 division, 2000 means 2%
+
+#time estimation
+#size 9:  65  sec
+#size 11: 100 sec
+#size 13: 160 sec
+
+make bp_decoding4.out
+
+stabilizer_folder=data/toric/stabilizer
+error_folder=data/toric/bp_decoding4
+#error_folder=data/toric/temp
+#test fold
+
+#size=13
+#for size in {5,7,9,11,13}
+for size in {13,11,9,7,5}
+#for size in 5
+do
+    
+    echo start run_save_bp.sh for size ${size}
+    date
+    for i in {100..5000..100}
+	     #for i in {2000..2000..100}
+    do
+    #add G matrix on Dec 14
+	./bp_decoding4.out ${stabilizer_folder}/toric_S_x_size_${size}.mm ${stabilizer_folder}/toric_S_z_size_${size}.mm ${error_folder}/toric_S_size_${size}.mm_rate $i &
+	echo $i
+	num_process=`pgrep -c decoding`
+	while (( num_process > 15 ))
+	do
+	    sleep 1
+	    num_process=`pgrep -c decoding`
+	done
+    done
+
+    #wait
+    echo start \(finish\) bp_decoding 10000 cycle for size ${size} x ${size} when p is \in 100,000 division. All error saved
+    date
+done
+wait
+
+date
+echo done
+#echo finish bp_decoding 10000 cycle for size ${size} x ${size} when p is \in 100,000 division. All error saved>> data/toric/bp_converge3/bp_decoding.log &
+#date >> data/toric/bp_converge3/bp_decoding.log &
diff --git a/temp/async_test.c b/temp/async_test.c
index e131688..9d82c75 100644
--- a/temp/async_test.c
+++ b/temp/async_test.c
@@ -4,9 +4,7 @@
 
 // a non-optimized way of checking for prime numbers:
 
- 
-
-	  
+ 	  
 bool is_prime (int x) {
   std::cout << "wait..."<<x<<"\n";
   for ( int i = 0; i<x; i++) double a=100*15-1/4.1+133.1/2.77/3.88*i+1.3*i;
@@ -32,7 +30,7 @@ int main ()
   int n = 40;
   std::future<int> fut[n];
   for ( int i =0;i<n;i++){
-    fut[i] = std::async (std::launch::async,get_prime,871782918);
+    fut[i] = std::async (std::launch::async,get_prime,71782918);
   }
 
   int num = fut3.get();
diff --git a/temp/async_test.out b/temp/async_test.out
index 9fde94a..8e9cff2 100755
Binary files a/temp/async_test.out and b/temp/async_test.out differ
diff --git a/temp/makefile~ b/temp/makefile~
deleted file mode 100644
index 879f515..0000000
--- a/temp/makefile~
+++ /dev/null
@@ -1,2 +0,0 @@
-async_test.out:async_test.c
-	g++ -o async_test.out async_test.c -std=c++11 -pthread
diff --git a/temp/async_test.c~ b/temp/prime_calculator.c
similarity index 90%
rename from temp/async_test.c~
rename to temp/prime_calculator.c
index 61b92f4..9d82c75 100644
--- a/temp/async_test.c~
+++ b/temp/prime_calculator.c
@@ -4,9 +4,7 @@
 
 // a non-optimized way of checking for prime numbers:
 
- 
-
-	  
+ 	  
 bool is_prime (int x) {
   std::cout << "wait..."<<x<<"\n";
   for ( int i = 0; i<x; i++) double a=100*15-1/4.1+133.1/2.77/3.88*i+1.3*i;
@@ -29,9 +27,10 @@ int main ()
   //std::future<bool> fut2 = std::async (is_prime,87178291199);
   std::future<int> fut3 = std::async (std::launch::async,get_prime,871782922);
   std::future<int> fut4 = std::async (std::launch::async,get_prime,871782920);
-  std::future<int> fut[10];
-  for ( int i =0;i<10;i++){
-    fut[i] = std::async (std::launch::async,get_prime,871782918);
+  int n = 40;
+  std::future<int> fut[n];
+  for ( int i =0;i<n;i++){
+    fut[i] = std::async (std::launch::async,get_prime,71782918);
   }
 
   int num = fut3.get();
diff --git a/test.c~ b/test.c~
deleted file mode 100644
index 8cac13f..0000000
--- a/test.c~
+++ /dev/null
@@ -1,63 +0,0 @@
-
-// CPP program to demonstrate multithreading
-// using three different callables.
-#include <iostream>
-#include <thread>
-using namespace std;
-
-// A dummy function
-void foo(int Z)
-{
-  for (int i = 0; i < Z; i++) {
-            cout << "Thread using function"
-	      " pointer as callable\n";
-  }
-}
-
-// A callable object
-class thread_obj {
- public:
-  void operator()(int x)
-  {
-    for (int i = 0; i < x; i++)
-                  cout << "Thread using function"
-		    " object as  callable\n";
-  }
-};
-
-int main()
-{
-      cout << "Threads 1 and 2 and 3 "
-	"operating independently" << endl;
-
-      // This thread is launched by using
-      // function pointer as callable
-      thread th1(foo, 3);
-
-      // This thread is launched by using
-      // function object as callable
-      thread th2(thread_obj(), 3);
-
-      // Define a Lambda Expression
-      auto f = [](int x) {
-	for (int i = 0; i < x; i++)
-	              cout << "Thread using lambda"
-			" expression as callable\n";
-      };
-
-      // This thread is launched by using
-      // lamda expression as callable
-      thread th3(f, 3);
-
-      // Wait for the threads to finish
-      // Wait for thread t1 to finish
-      th1.join();
-
-      // Wait for thread t2 to finish
-      th2.join();
-
-      // Wait for thread t3 to finish
-      th3.join();
-
-      return 0;
-}