try to do kernel decomp

shaders/A2Task1KernelDecomposition.comp

@@ -11,14 +11,36 @@ layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 layout(push_constant) uniform PushStruct {
     uint size;
-    uint offset;
 } p;
 
-layout(binding = 0) buffer inBuffer { uint v[]; };
-layout(binding = 1) buffer outBuffer { uint g_v[]; };
+layout(binding = 0) buffer inBuffer {
+    uint v[];
+};
+layout(binding = 1) buffer outBuffer {
+    uint g_v[];
+};
 
 // TODO: Shared variables
+// 512 Elements but initial reduction is done
+const uint bufferSize = 256;
+shared uint[bufferSize] localBuffer;
 
 void main() {
     // TODO: Kernel implementation
+
+    for (uint i = p.size / 2; i < 0; i -= 2) {
+        localBuffer[i] = v[i] + v[i + 1];
+    }
+
+    for (uint j = bufferSize ; j != 0; j / 2) {
+        for (uint i = bufferSize / 2; i < 0; i -= 2) {
+            localBuffer[i] = localBuffer[i] + localBuffer[i + 1];
+        }
+    }
+
+    localBuffer[0] = localBuffer[0] + localBuffer[1];
+
+    for (uint i = 0; i < bufferSize; i ++) {
+        g_v[i] = localBuffer[i];
+    }
 }

shaders/A2Task1Sequential.comp

@@ -19,9 +19,5 @@ layout(binding = 0) buffer inoutBufer { uint v[]; };
 void main() {    
     uint gIDx = gl_GlobalInvocationID.x;
     
-    if (gIDx + p.offset >= p.size) {
-        return;
-    }
-
     v[gIDx] += v[gIDx + p.offset];
 }