我想将A X乘以B。因此，（A X）+ B。但是，它不起作用，另外我需要在A的结果中找到最大数A是5 * 3矩阵，B是3的数组，C也是5个数字的数组
我问如何解决该错误以及我的代码中有什么错误另外，我无法使代码在矩阵A
的结果中找到最大数错误是这样的：PC = 0x00400084发生异常数据/堆栈中的错误地址读取：0x10040100PC = 0x00400084发生异常数据/堆栈中的错误地址读取：0x10040114
la $s4, matrixA #s4 set to base address of matrixA
la $s3, vectorX #s3 set to base address of vectorX
la $s7, vectorB #s7 set to base address of vectorB
la $s5, shapeA #s5 set to base address of sizeA
nop
lw $s6, 4($s5) #s6 set to second val in sizeA (col#)
nop
lw $s5, 0($s5) #s5 set to first val in sizeA (row#)
lw $t1, 0($s5) # $t1 = 32 (row size/loop end)
li $s0, 0 # i = 0; initialize 1st for loop
L1: li $s1,0 #j=0;restart2ndforloop
L2: li $s2,0 #k=0;restart3rdforloop
mul $t2, $s0, $s5 #$t2 = A[row](size of row of B)
addiu $t2, $t2, 0 # $t2 = i * size(row) + 0
sll $t2, $t2, 2 # $t2 = byte offset of [i][0]
addu $t2, $s7, $t2 # $t2 = byte address of B[i][0]
l.d $f4, 0($t2) # $f4 = 4 bytes of B[i][0]
L3: mul $t0,$s2,$s5 #$t0 = i * A[col](size of row of X)
addiu $t0, $t0, 0 # $t0 = k * size(row) + 0
sll $t0, $t0, 2 # $t0 = byte offset of X[k][0]
addu $t0, $s3, $t0 # $t0 = byte address of X[k][0]
l.d $f16, 0($t0) # $f16 = 4 bytes of X[k][0]
mul $t0,$s0,$s5 #$t0 = i * A[row](size of row of A)
addu $t0, $t0, $s2 # $t0 = i * size(row) + k
sll $t0, $t0, 2 # $t0 = byte offset of A[i][k]
addu $t0, $s4, $t0 # $t0 = byte address of A[i][k]
l.d $f18, 0($t0) # $f18 = 4 bytes of A[i][k]
mul.d $f16, $f18, $f16 # $f16 = y[i][k] * z[k][j]
add.d $f4, $f4, $f16 # $f4=x[i][j] + y[i][k]*z[k][j]
addiu $s2, $s2, 1 # k = k + 1
bne $s2,$t1,L3
s.d $f4, 0($t2)
addiu $s2, $s2, 1 # j = j + 1
bne $s2,$t1,L2
addiu $s0, $s0, 1 # i = i + 1
bne $s0,$t1,L1 #if(i!=32)gotoL1
我绝对不理解您的代码。您声明要进行矩阵矢量乘法，但是在注释中有“ x [i] [j] + y [i] [k] * z [k] [j]”。当我看到地址计算时，我觉得您正在处理浮点数，但是在代码中，它似乎与双精度有关。...
矩阵矢量乘法不需要复杂的地址计算，可以通过仅增加指针来完成。
您有兴趣在代码中分离循环计数和地址计算，并反转循环。
C代码b [] + = a [] [] * x []
for(i=0;i<N;i++){
double r=b[i];
for(j=0;j<M;j++){
r+=a[i][j]*x[j];
}
b[i]=r;
}
可以用指针并通过反转循环来重写
pb=&b[0] ;
pa=&a[0][0];
for (i=N;i>0;i--){
px=&x[0];
r=*pb;
for (j=M;j>0;j--){
r = r + *px * *pa;
px++;
pa++;
}
*pb=r ;
pb++ ;
}
和mips版本，即后者C代码的直接翻译
la $s1, matrixA #s1 set to base address of matrixA pa
la $s2, vectorB #s2 set to base address of vectorB pb
la $s3, vectorX #s3 set to base address of vectorX px
# s4 temps for the vector x
la $s5, shapeA #s5 set to base address of sizeA
lw $s6, 4($s5) #s6 set to second val in sizeA (col#)
lw $s5, 0($s5) #s5 set to first val in sizeA (row#)
L1: li $t1,$s5 # i=M
move $s4,$s3 # $s4 == px
l.d $f0,0($s2) # f0=r=*pb
L2: li $t2,$s6 # j=N
l.d $f2,0($s1) # $f2 *pa=a[i]][j]
l.d $f4,0($s4) # $f4 *px=x[j]
mul.d $f2,$f2,$f4 # a[i][j]*x[j]
add.d $f0,$f0,$f2 # +b[i]
addi $t2,$t2,-1 # j--
addi $s1,$s1,8 # pa++
addi $s4,$s4,8 # px++
bne $t2,L2 # j>0?->L2
s.d $f0,0($s2) # *pb=r
addi $s2,$s2,8 # pb++
addi $t1,$t1,-1 # i--
bne $t1,L1 # j>0?->L1
main：
la $s1, matrixA #s1 set to base address of matrixA pa
la $s2, vectorB #s2 set to base address of vectorB pb
la $s3, vectorX #s3 set to base address of vectorX px
# s4 temps for the vector x
la $s5, shapeA #s5 set to base address of sizeA
lw $s6, 4($s5) #s6 set to second val in sizeA (col#)
lw $s5, 0($s5) #s5 set to first val in sizeA (row#)
L1: lw $t1,0($s5) # i=M
move $s4,$s3 # $s4 == px
l.d $f0,0($s2) # f0=r=*pb
L2: lw $t2,4($s5) # j=N
l.d $f2,0($s1) # $f2 *pa=a[i]][j]
l.d $f4,0($s4) # $f4 *px=x[j]
mul.d $f2,$f2,$f4 # a[i][j]*x[j]
add.d $f0,$f0,$f2 # +b[i]
addi $t2,$t2,-1 # j--
addi $s1,$s1,8 # pa++
addi $s4,$s4,8 # px++
slt $t5,$zero,$t2
bne $t5,$zero,L2 # j>0?->L2
s.d $f0,0($s2) # *pb=r
addi $s2,$s2,8 # pb++
addi $t1,$t1,-1 # i--
slt $t5,$zero,$t1
bne $t5,$zero,L1 # i>0?->L1
# print the values of y,
# where y = max(0, matrixA * vectorX + vectorB)
li $v0, 4 # system call code for printing string = 4
la $a0, str1
syscall # call operating system to perform operation
# printing shape
li $v0, 1
la $s0, shapeA
lw $a0, 0($s0)
syscall
li $v0, 4 # system call code for printing string = 4
la $a0, space
syscall # call operating system to perform operation
li $v0, 1
lw $a0, 4($s0)
syscall
li $v0, 4 # system call code for printing string = 4
la $a0, str2
syscall # call operating system to perform operation
# printing the first two elements
li $v0, 2
la $s1, matrixA
lwc1 $f12, 0($s1)
syscall
li $v0, 4 # system call code for printing string = 4
la $a0, space
syscall # call operating system to perform operation
li $v0, 2
lwc1 $f12, 4($s1)
syscall
我这样更改了它，但它正在产生异常错误