saxpy test: rs1 error

rvvector/benchmark_temp/Makefile

@@ -34,8 +34,8 @@ HEX: ELF
 	$(CPY) -O ihex $(VX_MAIN).elf $(VX_MAIN).hex
 
 ELF:
-	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC1) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
+#	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC1) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
-#	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC2) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
+	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC2) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
 #	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC3) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
 #	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC4) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf
 #	$(COMP) $(CC_FLAGS) $(VX_STR) $(VX_VEC5) $(VX_FIO) $(NEWLIB) $(VX_INT) $(VX_IO) $(VX_API) $(VX_TEST) $(VX_MAIN).c $(LIBS) -Iinclude  -o $(VX_MAIN).elf~                                

rvvector/benchmark_temp/vx_vec_benchmark.c

@@ -6,14 +6,17 @@
 int main()
 {
     vx_tmc(1);
+
     int n = 5;
+    int scalar = 10;
+
     int *a = (int*)malloc(sizeof(int) * n); //{1, 1, 1, 1, 1};
     int *b = (int*)malloc(sizeof(int) * n); //{1, 1, 1, 1, 1};
     int *c = (int*)malloc(sizeof(int) * n); //{1, 1, 1, 1, 1};
 
     for (int i = 0; i < n; ++i) { a[i] = 1; b[i] = 2; c[i] = 5; }
 
-#if 1
+#if 0
 //---------------------------------------------------------------
 /* vvaddint32
  * # vector-vector add routine of 32-bit integers
@@ -43,7 +46,6 @@ int main()
 /* #  vector-scalar add
    # for (i=0; i<N; i++) { C[i] = A[i] + B; } // 32-bit ints */
     for (int i = 0; i < n; ++i) { a[i] = 1; b[i] = 1;}
-    int scalar = 10;
     printf("vsadd...scalar:%d\na[%d]: ", scalar, n);
     for(int i = 0; i < n; ++i) printf("%d \n", a[i]);
     printf("\nb: %d", scalar);
@@ -78,10 +80,12 @@ int main()
         if(a[i] != b[i]) 
         {
            printf("\n<memcpy> failed at <index: %d>! \n", i);
-           return;   
+           return 1;   
         }
     }
     printf("\nPASSED.......................... <memcpy> \n");
+#endif
+#if 1
 //---------------------------------------------------------------
 /* # void saxpy(size_t n, const float a, const float *x, float *y)
    # ==> convert to int!!
@@ -99,16 +103,22 @@ int main()
 
     vx_vec_saxpy(n, scalar, a, b);
 
+    printf("saxpy\na[%d]: ", n);
+    for(int i = 0; i < n; ++i) printf("%d \n", a[i]);
+    printf("\nb[%d]: ", n);
+    for(int i = 0; i < n; ++i) printf("%d \n", b[i]);
+
     for(int i = 0; i < n; ++i) 
     {
         if(b[i] != ((a[i] * scalar) + c[i])) 
         {
            printf("\n<saxpy> failed at <index: %d>! \n", i);
-           return;   
+           return 1;   
         }
     }
     printf("\nPASSED.......................... <saxpy> \n");
-
+#endif
+#if 0
 //---------------------------------------------------------------
 /* # void sgemm_nn(size_t n, size_t m, size_t k, const float*a,   // m * k matrix
 #          size_t lda, const float*b,   // k * n matrix 

rvvector/benchmark_temp/vx_vec_benchmark.h

@@ -5,10 +5,10 @@
 extern "C" {
 #endif
 
-void vx_vec_vvaddint32(int n, int* a, int* b, int *c);
+//void vx_vec_vvaddint32(int n, int* a, int* b, int *c);
 //void vx_vec_vsadd(int n, int* a, int scalar);
 //void vx_vec_memcpy(int* a, int* b, int n);
-//void vx_vec_saxpy(int n, int scalar, int* a, int* b);
+void vx_vec_saxpy(int n, int scalar, int* a, int* b);
 //void vx_vec_sgemm_nn(int n, int m, int k, int* a1, int lda, int* b1, int ldb, int* c1, int ldc);
 
 #ifdef __cplusplus

rvvector/benchmark_temp/vx_vec_saxpy.s

@@ -13,16 +13,31 @@
 #     fa0     a
 #     a1      x
 #     a2      y
+#vx_vec_saxpy:
+#    vsetvli a4, a0, e32, m8
+#saxpy:
+#    vlw.v v0, (a1)
+#    sub a0, a0, a4
+#    slli a4, a4, 2
+#    add a1, a1, a4
+#    vlw.v v8, (a2)
+#    vfmacc.vf v8, fa0, v0
+#    vsw.v v8, (a2)
+#    add a2, a2, a4
+#    bnez a0, saxpy
+#    ret
+
+# a0 n, rs1 a, a2 x, a3 y
 vx_vec_saxpy:
     vsetvli a4, a0, e32, m8
 saxpy:
-    vlw.v v0, (a1)
+    vlw.v v0, (a2)
     sub a0, a0, a4
     slli a4, a4, 2
-    add a1, a1, a4
-    vlw.v v8, (a2)
-    vfmacc.vf v8, fa0, v0
-    vsw.v v8, (a2)
     add a2, a2, a4
+    vlw.v v1, (a3)
+    vmacc.vx v1, rs1, v0
+    vsw.v v1, (a3)
+    add a3, a3, a4
     bnez a0, saxpy
     ret