Venus RISC-V 如何循环、比较和打印？

Question

I am trying to loop through an array and if the number is larger than X then print.

I've tried to find tutorials online but I'm just stuck on why it is not working/outputting anything. My comments kind of explain what i tried to do.

.data
arrayOfNums:
.word 0
.word 1
.word 122
.word 1112
.word 4294967295
.word 22
.word 234234
.word 23332
.word 42
.word 23423

K: .word 2237

.text
.globl main

main:
####   *** vv My problem starts here vv ***   ####
la t0 K             #set t0 to K
la t1 arrayOfNums   #set t1 pointer to array
addi a1 t0 0        #add t0 deallocated to a1
addi a2 t1 0        #add t1 deallocated to a2

loop: 
    addi a0 x0 1        # print_int ecall
    addi a1 t1 0        # add t1 value to print
    ecall
    addi t1, t1, 4      # Increments t1 to move to the next element
    jal x0, loop

exit:
####Exit using environmental calls####
addi a1 x0 0         # random exit 0 
addi a0 x0 17        # print_int ecall
ecall

Thank you!

Answer 1

There are some issues with the code you have posted.

I am trying to loop through an array and if the number is larger than X then print.

I can't find any X symbol in your code. Do you mean K ?

 la t0 K #set t0 to K

The comment is wrong. You load the address of K into t0 . If you want to get the value that is stored at that address into t0 you have to load the address into another register and dereference that address into t0 , ie load it with the lw or lwu instruction.

 addi a1 t0 0 #add t0 deallocated to a1 addi a2 t1 0 #add t1 deallocated to a2

What do you mean with 'deallocated to'? You copy t0 to a1 and t1 to a2 . The same could be archieved with the mv pseudo instruction. However, those 2 lines are superfluous as you immediately overwrite the a1 and a2 registers in the following lines.

 addi a0 x0 1 # print_int ecall addi a1 t1 0 # add t1 value to print ecall

You could use the li / mv pseudo instructions here. This unconditionally prints the value of t1 - which is an address. If you want to print an actual array element you would have to load it using the address stored in t1 - cf. the lw / lwu instructions.

 addi t1, t1, 4 # Increments t1 to move to the next element jal x0, loop

With that you unconditionally jump to the head of the loop (with pseudo instruction: j loop ), ie that means that you read over the end of your array and never quit the loop. To fix this you have to use a conditional branch instruction such as bnez . Meaning that you eg set a register (as counter) to the array size and decrement it until zero. Or set a register to the address after the last array element and branch until t1 is equal to it.

 addi a1 x0 0 # random exit 0 addi a0 x0 17 # print_int ecall ecall

It's anything but random. The comment is incorrect, you are calling the Venus exit2 syscall, not print_int. Besides, Venus also provides an exit syscall ( 10 ) that doesn't require an argument.

What is completely missing from your code is a location where you are actually trying to compare numbers and then print them based on that comparison.