# Assignment2: RISC-V Toolchain
## Problem Description
The problem is LeetCode136 from [吳紀寬](https://hackmd.io/JqxPI8fpSP-_0JBFB2t-Tw?view).
The following is the Description
>Given a non-empty array of integers nums, every element appears twice except for one. Find that single one.
>You must implement a solution with a linear runtime complexity and use only constant extra space.
## Motivation
This problem seems simple at first. It can be implemented with multiple answer, but [吳紀寬](https://hackmd.io/JqxPI8fpSP-_0JBFB2t-Tw?view)'s answer looks interesting. I don't know that the XOR can use by this way. In my own answer, it need an extra array to record the number. But, this way will use more memory, and his answer looks more easy and efficient.
## C Code
[吳紀寬](https://hackmd.io/JqxPI8fpSP-_0JBFB2t-Tw?view)'s C code is perfect already. I just move the ```singlenumber``` function to the head, making it can run in DevC++.
Because rv32emu print a number under 10 is more easier, so I change the arr2 test data from```1, 2, 3, 1, 3, 2, *12* ```to```1, 2, 3, 1, 3, 2, *5* ```. Hence the answer will be changed from 12 to 5.
```clike=
#include<stdio.h>
int singlenumber(int* arr, int size)
{
int res = arr[0];
for (int i = 1; i < size; i++)
res ^= arr[i];
return res;
}
int main() {
int arr1[] = { 4, 1, 2, 1, 2 };
int size1 = 5;
int res1 = singlenumber(arr1, size1);
printf("result 1: %d \n", res1);
int arr2[] = { 1, 2, 3, 1, 3, 2, 5 };
int size2 = 7;
int res2 = singlenumber(arr2, size2);
printf("result 2: %d \n", res2);
int arr3[] = { 3, 6, 3, 6, 7, 8, 8, 5, 5 };
int size3 = 9;
int res3 = singlenumber(arr3, size3);
printf("result 3: %d \n", res3);
}
```
## Assembly(Origin)
``` assembly=
.data
arr1: .word 4, 1, 2, 1, 2
arr2: .word 1, 2, 3, 1, 3, 2, 5
arr3: .word 3, 6, 3, 6, 7, 8, 8, 5, 5
size1: .word 5
size2: .word 7
size3: .word 9
result1: .string "result 1: "
result2: .string "result 2: "
result3: .string "result 3: "
space: .string " "
enter: .string "\n"
.text
main:
la a1, arr1 #load base address of arr1 into a1
lw a2, size1 #load size of arr1 into a2
jal ra, printarray #call for print example array function
la a0, result1 #print string result1
li a7, 4
ecall
jal ra, singlenumber #call for singlenumber function
jal Enter
la a1, arr2 #load base address of arr2 into a1
lw a2, size2 #load size of arr2 into a2
jal ra, printarray #call for print example array function
la a0, result2 #print string result2
li a7, 4
ecall
jal ra, singlenumber #call for singlenumber function
jal Enter #print string enter
la a1, arr3 #load base address of arr3 into a1
lw a2, size3 #load size of arr3 into a2
jal ra, printarray #call for print example array function
la a0, result3 #print string result3
li a7, 4
ecall
jal ra, singlenumber #call for singlenumber function
jal Enter #print string enter
li a7 10
ecall
singlenumber:
mv t0, x0 #t0 as temporary result
mv t1, x0 #t1 as i of for loop
mv t2, a1 #t2 as temporay address and initialize with address of example array
loop1:
beq t1, a2, exit1 #if i==size, exit1
lw t3, 0(t2) #load array element into t3
xor t0, t0, t3 #xor operation, and save result into t0
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
j loop1 #jump back to loop1
printarray:
mv t1, x0 #t1 as i of for loop
mv t2, a1 #t2 as temporay address and initialize with address of example array
loop2:
beq t1, a2, exit2 #if i == size, exit2
lw a0, 0(t2) #load array element into a0
li a7, 1 #print array element
ecall
la a0, space #print space
li a7, 4
ecall
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
j loop2 #jump back to loop2
exit1:
mv a0, t0 #print result in a0
li a7, 1
ecall
jr ra #jump back to main function
exit2:
la a0, enter #print enter
li a7, 4
ecall
jr ra #jump back to main function
Enter:
la a0, enter #print enter
li a7, 4
ecall
jr ra #jump back to main function
```
## After
Because the three function exit, exit2 and Enter only print the ```\n``` string. The exit and Enter is the same. I combine the three functions into previous function. By this way, I can remove some jal and jr and make the code shortly .
In loop1 and loop2, I change the beq to bne and remove the jump instruction and add print function(print answer and ```\n```) behind. Meanwhile, the main can remove the print function and jump instruction. Furthermore, because some ```li``` overlap, I can remove some unnecessary ```li```, too.
```assembly=
.data
arr1: .word 4, 1, 2, 1, 2
arr2: .word 1, 2, 3, 1, 3, 2, 5
arr3: .word 3, 6, 3, 6, 7, 8, 8, 5, 5
size1: .word 5
size2: .word 7
size3: .word 9
result1: .string "result 1: "
result2: .string "result 2: "
result3: .string "result 3: "
space: .string " "
enter: .string "\n"
.text
main:
la a1, arr1 #load base address of arr1 into a1
lw a2, size1 #load size of arr1 into a2
jal ra, printarray #call for print example array function
la a0, result1 #print string result1
ecall
jal ra, singlenumber #call for singlenumber function
la a1, arr2 #load base address of arr2 into a1
lw a2, size2 #load size of arr2 into a2
jal ra, printarray #call for print example array function
la a0, result2 #print string result2
ecall
jal ra, singlenumber #call for singlenumber function
la a1, arr3 #load base address of arr3 into a1
lw a2, size3 #load size of arr3 into a2
jal ra, printarray #call for print example array function
la a0, result3 #print string result3
ecall
jal ra, singlenumber #call for singlenumber function #print string enter
li a7 10
ecall
singlenumber:
mv t0, x0 #t0 as temporary result
mv t1, x0 #t1 as i of for loop
mv t2, a1 #t2 as temporay address and initialize with address of example array
loop1:
lw t3, 0(t2) #load array element into t3
xor t0, t0, t3 #xor operation, and save result into t0
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
bne t1, a2, loop1 #if t1<a2, jump back to loop1
mv a0, t0 #print result in a0
li a7, 1
ecall
la a0, enter #print enter
li a7, 4
ecall
jr ra
printarray:
mv t1, x0 #t1 as i of for loop
mv t2, a1 #t2 as temporay address and initialize with address of example array
loop2:
lw a0, 0(t2) #load array element into a0
li a7, 1 #print array element
ecall
la a0, space #print space
li a7, 4
ecall
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
bne t1, a2, loop2 #if i == size, exit2
la a0, enter #print enter
ecall
jr ra #jump back to main function
```
From the picture above, the total cycle reduced.
## Convert to RV32I
I change the print way to make rv32emu can run.
``` assembly=
.org 0
.global _start
/* newlib system calls */
.set SYSEXIT, 93
.set SYSWRITE, 64
.data
arr1: .word 4, 1, 2, 1, 2
.set arr_size, .-arr1
arr2: .word 1, 2, 3, 1, 3, 2, 5
.set arr_size, .-arr2
arr3: .word 3, 6, 3, 6, 7, 8, 8, 5, 5
.set arr_size, .-arr3
size1: .word 5
size2: .word 7
size3: .word 9
result1: .ascii "result 1: "
.set str_result1_len, .-result1
result2: .ascii "result 2: "
.set str_result2_len, .-result2
result3: .ascii "result 3: "
.set str_result3_len, .-result3
space: .ascii " "
.set str_space_len, .-space
enter: .ascii "\n"
.set str_entere_len, .-enter
.text
_start:
la s1, arr1 #load base address of arr1 into s1
lw s2, size1 #load size of arr1 into s2
jal ra, printarray
la a1, result1 #print string result1
li a0 , 1
li a2, str_result1_len
li a7, SYSWRITE
ecall
jal ra, singlenumber #call for singlenumber function
la s1, arr2 #load base address of arr2 into s1
lw s2, size2 #load size of arr2 into s2
jal ra, printarray
la a1, result2 #print string result2
li a0 , 1
li a2, str_result2_len
li a7, SYSWRITE
ecall
jal ra, singlenumber #call for singlenumber function
la s1, arr3 #load base address of arr3 into s1
lw s2, size3 #load size of arr3 into s2
jal ra, printarray
la a1, result3 #print string result3
li a0 , 1
li a2, str_result3_len
li a7, SYSWRITE
ecall
jal ra, singlenumber #call for singlenumber function
li a0, 0
li a7, SYSEXIT
ecall
singlenumber:
mv t0, x0 #t0 as temporary result
mv t1, x0 #t1 as i of for loop
mv t2, s1 #t2 as temporay address and initialize with address of example array
loop1:
lw t3, 0(t2) #load array element into t3
xor t0, t0, t3 #xor operation, and save result into t0
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
bne t1, s2, loop1 #if t1<s2, jump back to loop1
mv a1, t0 #print result in t0
addi a1,a1,48 #to ASCII
addi sp, sp, -4
sw a1, 0(sp)
addi a1, sp, 0
li a0 , 1
li a2, 4
li a7, SYSWRITE
ecall
addi sp,sp,4
la a1, enter #print enter
li a0 , 1
li a2, str_entere_len
li a7, SYSWRITE
ecall
jr ra
printarray:
mv t1, x0 #t1 as i of for loop
mv t2, s1 #t2 as temporay address and initialize with address of example array
loop2:
lw a1, 0(t2) #load array element into a0
addi a1,a1,48
addi sp, sp, -4
sw a1, 0(sp)
addi a1, sp, 0
li a0 , 1
li a2, 4
li a7, SYSWRITE
ecall
addi sp,sp,4
la a1, space #print space
li a0 , 1
li a2, str_space_len
li a7, SYSWRITE
ecall
addi t2, t2, 4 #shift address by 1 word
addi t1, t1, 1 #add i by 1
bne t1, s2, loop2 #if i == size, exit2
la a1, enter #print enter
li a0 , 1
li a2, str_entere_len
li a7, SYSWRITE
ecall
jr ra
```
```
4 1 2 1 2
result 1: 4
1 2 3 1 3 2 5
result 2: 5
3 6 3 6 7 8 8 5 5
result 3: 7
inferior exit code 0
CSR cycle count: 576
```
---
## -O0
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 75104 | 2816 |812 |78732 |1338c |SingleNumber_o0 |
### Header
### Assembly code after compiling
``` clike=
00010184 <singlenumber>:
10184: fd010113 addi sp,sp,-48
10188: 02812623 sw s0,44(sp)
1018c: 03010413 addi s0,sp,48
10190: fca42e23 sw a0,-36(s0)
10194: fcb42c23 sw a1,-40(s0)
10198: fdc42783 lw a5,-36(s0)
1019c: 0007a783 lw a5,0(a5)
101a0: fef42623 sw a5,-20(s0)
101a4: 00100793 li a5,1
101a8: fef42423 sw a5,-24(s0)
101ac: 0300006f j 101dc <singlenumber+0x58>
101b0: fe842783 lw a5,-24(s0)
101b4: 00279793 slli a5,a5,0x2
101b8: fdc42703 lw a4,-36(s0)
101bc: 00f707b3 add a5,a4,a5
101c0: 0007a783 lw a5,0(a5)
101c4: fec42703 lw a4,-20(s0)
101c8: 00f747b3 xor a5,a4,a5
101cc: fef42623 sw a5,-20(s0)
101d0: fe842783 lw a5,-24(s0)
101d4: 00178793 addi a5,a5,1
101d8: fef42423 sw a5,-24(s0)
101dc: fe842703 lw a4,-24(s0)
101e0: fd842783 lw a5,-40(s0)
101e4: fcf746e3 blt a4,a5,101b0 <singlenumber+0x2c>
101e8: fec42783 lw a5,-20(s0)
101ec: 00078513 mv a0,a5
101f0: 02c12403 lw s0,44(sp)
101f4: 03010113 addi sp,sp,48
101f8: 00008067 ret
000101fc <main>:
101fc: f7010113 addi sp,sp,-144
10200: 08112623 sw ra,140(sp)
10204: 08812423 sw s0,136(sp)
10208: 09010413 addi s0,sp,144
1020c: f6a42e23 sw a0,-132(s0)
10210: f6b42c23 sw a1,-136(s0)
10214: 000227b7 lui a5,0x22
10218: 80878793 addi a5,a5,-2040 # 21808 <__clzsi2+0xbc>
1021c: 0007a583 lw a1,0(a5)
10220: 0047a603 lw a2,4(a5)
10224: 0087a683 lw a3,8(a5)
10228: 00c7a703 lw a4,12(a5)
1022c: 0107a783 lw a5,16(a5)
10230: fcb42223 sw a1,-60(s0)
10234: fcc42423 sw a2,-56(s0)
10238: fcd42623 sw a3,-52(s0)
1023c: fce42823 sw a4,-48(s0)
10240: fcf42a23 sw a5,-44(s0)
10244: 00500793 li a5,5
10248: fef42623 sw a5,-20(s0)
1024c: fc440793 addi a5,s0,-60
10250: fec42583 lw a1,-20(s0)
10254: 00078513 mv a0,a5
10258: f2dff0ef jal ra,10184 <singlenumber>
1025c: fea42423 sw a0,-24(s0)
10260: fe842583 lw a1,-24(s0)
10264: 000217b7 lui a5,0x21
10268: 7d878513 addi a0,a5,2008 # 217d8 <__clzsi2+0x8c>
1026c: 324000ef jal ra,10590 <printf>
10270: 000227b7 lui a5,0x22
10274: 81c78793 addi a5,a5,-2020 # 2181c <__clzsi2+0xd0>
10278: 0007a803 lw a6,0(a5)
1027c: 0047a503 lw a0,4(a5)
10280: 0087a583 lw a1,8(a5)
10284: 00c7a603 lw a2,12(a5)
10288: 0107a683 lw a3,16(a5)
1028c: 0147a703 lw a4,20(a5)
10290: 0187a783 lw a5,24(a5)
10294: fb042423 sw a6,-88(s0)
10298: faa42623 sw a0,-84(s0)
1029c: fab42823 sw a1,-80(s0)
102a0: fac42a23 sw a2,-76(s0)
102a4: fad42c23 sw a3,-72(s0)
102a8: fae42e23 sw a4,-68(s0)
102ac: fcf42023 sw a5,-64(s0)
102b0: 00700793 li a5,7
102b4: fef42223 sw a5,-28(s0)
102b8: fa840793 addi a5,s0,-88
102bc: fe442583 lw a1,-28(s0)
102c0: 00078513 mv a0,a5
102c4: ec1ff0ef jal ra,10184 <singlenumber>
102c8: fea42023 sw a0,-32(s0)
102cc: fe042583 lw a1,-32(s0)
102d0: 000217b7 lui a5,0x21
102d4: 7e878513 addi a0,a5,2024 # 217e8 <__clzsi2+0x9c>
102d8: 2b8000ef jal ra,10590 <printf>
102dc: 000227b7 lui a5,0x22
102e0: 83878793 addi a5,a5,-1992 # 21838 <__clzsi2+0xec>
102e4: 0007a303 lw t1,0(a5)
102e8: 0047a883 lw a7,4(a5)
102ec: 0087a803 lw a6,8(a5)
102f0: 00c7a503 lw a0,12(a5)
102f4: 0107a583 lw a1,16(a5)
102f8: 0147a603 lw a2,20(a5)
102fc: 0187a683 lw a3,24(a5)
10300: 01c7a703 lw a4,28(a5)
10304: 0207a783 lw a5,32(a5)
10308: f8642223 sw t1,-124(s0)
1030c: f9142423 sw a7,-120(s0)
10310: f9042623 sw a6,-116(s0)
10314: f8a42823 sw a0,-112(s0)
10318: f8b42a23 sw a1,-108(s0)
1031c: f8c42c23 sw a2,-104(s0)
10320: f8d42e23 sw a3,-100(s0)
10324: fae42023 sw a4,-96(s0)
10328: faf42223 sw a5,-92(s0)
1032c: 00900793 li a5,9
10330: fcf42e23 sw a5,-36(s0)
10334: f8440793 addi a5,s0,-124
10338: fdc42583 lw a1,-36(s0)
1033c: 00078513 mv a0,a5
10340: e45ff0ef jal ra,10184 <singlenumber>
10344: fca42c23 sw a0,-40(s0)
10348: fd842583 lw a1,-40(s0)
1034c: 000217b7 lui a5,0x21
10350: 7f878513 addi a0,a5,2040 # 217f8 <__clzsi2+0xac>
10354: 23c000ef jal ra,10590 <printf>
10358: 00000793 li a5,0
1035c: 00078513 mv a0,a5
10360: 08c12083 lw ra,140(sp)
10364: 08812403 lw s0,136(sp)
10368: 09010113 addi sp,sp,144
1036c: 00008067 ret
```
### Observation
* Line of code: ```126```
* Allocate ```144``` bytes on stack
* Registers used: ```ra``` ```sp``` ```t1``` ```a0 ~ a7``` ```s0 ```
* Number of lw and sw : ```lw: 40``` ```sw: 38 ```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4792
```
## -O1
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 74960 | 2816 |812 |78588 |132fc |SingleNumber_o1 |
### Header
### Assembly code after compiling
``` clike=
00010184 <singlenumber>:
10184: fd010113 addi sp,sp,-48
10188: 02812623 sw s0,44(sp)
1018c: 03010413 addi s0,sp,48
10190: fca42e23 sw a0,-36(s0)
10194: fcb42c23 sw a1,-40(s0)
10198: fdc42783 lw a5,-36(s0)
1019c: 0007a783 lw a5,0(a5)
101a0: fef42623 sw a5,-20(s0)
101a4: 00100793 li a5,1
101a8: fef42423 sw a5,-24(s0)
101ac: 0300006f j 101dc <singlenumber+0x58>
101b0: fe842783 lw a5,-24(s0)
101b4: 00279793 slli a5,a5,0x2
101b8: fdc42703 lw a4,-36(s0)
101bc: 00f707b3 add a5,a4,a5
101c0: 0007a783 lw a5,0(a5)
101c4: fec42703 lw a4,-20(s0)
101c8: 00f747b3 xor a5,a4,a5
101cc: fef42623 sw a5,-20(s0)
101d0: fe842783 lw a5,-24(s0)
101d4: 00178793 addi a5,a5,1
101d8: fef42423 sw a5,-24(s0)
101dc: fe842703 lw a4,-24(s0)
101e0: fd842783 lw a5,-40(s0)
101e4: fcf746e3 blt a4,a5,101b0 <singlenumber+0x2c>
101e8: fec42783 lw a5,-20(s0)
101ec: 00078513 mv a0,a5
101f0: 02c12403 lw s0,44(sp)
101f4: 03010113 addi sp,sp,48
101f8: 00008067 ret
000101fc <main>:
101fc: f7010113 addi sp,sp,-144
10200: 08112623 sw ra,140(sp)
10204: 08812423 sw s0,136(sp)
10208: 09010413 addi s0,sp,144
1020c: f6a42e23 sw a0,-132(s0)
10210: f6b42c23 sw a1,-136(s0)
10214: 000227b7 lui a5,0x22
10218: 80878793 addi a5,a5,-2040 # 21808 <__clzsi2+0xbc>
1021c: 0007a583 lw a1,0(a5)
10220: 0047a603 lw a2,4(a5)
10224: 0087a683 lw a3,8(a5)
10228: 00c7a703 lw a4,12(a5)
1022c: 0107a783 lw a5,16(a5)
10230: fcb42223 sw a1,-60(s0)
10234: fcc42423 sw a2,-56(s0)
10238: fcd42623 sw a3,-52(s0)
1023c: fce42823 sw a4,-48(s0)
10240: fcf42a23 sw a5,-44(s0)
10244: 00500793 li a5,5
10248: fef42623 sw a5,-20(s0)
1024c: fc440793 addi a5,s0,-60
10250: fec42583 lw a1,-20(s0)
10254: 00078513 mv a0,a5
10258: f2dff0ef jal ra,10184 <singlenumber>
1025c: fea42423 sw a0,-24(s0)
10260: fe842583 lw a1,-24(s0)
10264: 000217b7 lui a5,0x21
10268: 7d878513 addi a0,a5,2008 # 217d8 <__clzsi2+0x8c>
1026c: 324000ef jal ra,10590 <printf>
10270: 000227b7 lui a5,0x22
10274: 81c78793 addi a5,a5,-2020 # 2181c <__clzsi2+0xd0>
10278: 0007a803 lw a6,0(a5)
1027c: 0047a503 lw a0,4(a5)
10280: 0087a583 lw a1,8(a5)
10284: 00c7a603 lw a2,12(a5)
10288: 0107a683 lw a3,16(a5)
1028c: 0147a703 lw a4,20(a5)
10290: 0187a783 lw a5,24(a5)
10294: fb042423 sw a6,-88(s0)
10298: faa42623 sw a0,-84(s0)
1029c: fab42823 sw a1,-80(s0)
102a0: fac42a23 sw a2,-76(s0)
102a4: fad42c23 sw a3,-72(s0)
102a8: fae42e23 sw a4,-68(s0)
102ac: fcf42023 sw a5,-64(s0)
102b0: 00700793 li a5,7
102b4: fef42223 sw a5,-28(s0)
102b8: fa840793 addi a5,s0,-88
102bc: fe442583 lw a1,-28(s0)
102c0: 00078513 mv a0,a5
102c4: ec1ff0ef jal ra,10184 <singlenumber>
102c8: fea42023 sw a0,-32(s0)
102cc: fe042583 lw a1,-32(s0)
102d0: 000217b7 lui a5,0x21
102d4: 7e878513 addi a0,a5,2024 # 217e8 <__clzsi2+0x9c>
102d8: 2b8000ef jal ra,10590 <printf>
102dc: 000227b7 lui a5,0x22
102e0: 83878793 addi a5,a5,-1992 # 21838 <__clzsi2+0xec>
102e4: 0007a303 lw t1,0(a5)
102e8: 0047a883 lw a7,4(a5)
102ec: 0087a803 lw a6,8(a5)
102f0: 00c7a503 lw a0,12(a5)
102f4: 0107a583 lw a1,16(a5)
102f8: 0147a603 lw a2,20(a5)
102fc: 0187a683 lw a3,24(a5)
10300: 01c7a703 lw a4,28(a5)
10304: 0207a783 lw a5,32(a5)
10308: f8642223 sw t1,-124(s0)
1030c: f9142423 sw a7,-120(s0)
10310: f9042623 sw a6,-116(s0)
10314: f8a42823 sw a0,-112(s0)
10318: f8b42a23 sw a1,-108(s0)
1031c: f8c42c23 sw a2,-104(s0)
10320: f8d42e23 sw a3,-100(s0)
10324: fae42023 sw a4,-96(s0)
10328: faf42223 sw a5,-92(s0)
1032c: 00900793 li a5,9
10330: fcf42e23 sw a5,-36(s0)
10334: f8440793 addi a5,s0,-124
10338: fdc42583 lw a1,-36(s0)
1033c: 00078513 mv a0,a5
10340: e45ff0ef jal ra,10184 <singlenumber>
10344: fca42c23 sw a0,-40(s0)
10348: fd842583 lw a1,-40(s0)
1034c: 000217b7 lui a5,0x21
10350: 7f878513 addi a0,a5,2040 # 217f8 <__clzsi2+0xac>
10354: 23c000ef jal ra,10590 <printf>
10358: 00000793 li a5,0
1035c: 00078513 mv a0,a5
10360: 08c12083 lw ra,140(sp)
10364: 08812403 lw s0,136(sp)
10368: 09010113 addi sp,sp,144
1036c: 00008067 ret
```
### Observation
* Line of code: ```126```
* Allocate ```144``` bytes on stack
* Registers used: ```ra``` ```sp``` ```a0 ~ a7``` ```s0 ```
* Number of lw and sw : ```lw: 40``` ```sw: 38 ```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4562
```
## -O2
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 74908 | 2816 |812 |78536 |132c8 |SingleNumber_o2 |
### Header
### Assembly code after compiling
``` clike=
000100c4 <main>:
100c4: fb010113 addi sp,sp,-80
100c8: 00021537 lui a0,0x21
100cc: 04812423 sw s0,72(sp)
100d0: 00400593 li a1,4
100d4: 72050513 addi a0,a0,1824 # 21720 <__clzsi2+0x88>
100d8: 00021437 lui s0,0x21
100dc: 75040413 addi s0,s0,1872 # 21750 <__clzsi2+0xb8>
100e0: 04112623 sw ra,76(sp)
100e4: 3f8000ef jal ra,104dc <printf>
100e8: 00842583 lw a1,8(s0)
100ec: 01042683 lw a3,16(s0)
100f0: 01842783 lw a5,24(s0)
100f4: 00042803 lw a6,0(s0)
100f8: 00442503 lw a0,4(s0)
100fc: 00c42603 lw a2,12(s0)
10100: 01442703 lw a4,20(s0)
10104: 00b12423 sw a1,8(sp)
10108: 00d12823 sw a3,16(sp)
1010c: 00f12c23 sw a5,24(sp)
10110: 01012023 sw a6,0(sp)
10114: 00a12223 sw a0,4(sp)
10118: 00c12623 sw a2,12(sp)
1011c: 00e12a23 sw a4,20(sp)
10120: 00010793 mv a5,sp
10124: 01810693 addi a3,sp,24
10128: 00100593 li a1,1
1012c: 0047a703 lw a4,4(a5)
10130: 00478793 addi a5,a5,4
10134: 00e5c5b3 xor a1,a1,a4
10138: fed79ae3 bne a5,a3,1012c <main+0x68>
1013c: 00021537 lui a0,0x21
10140: 73050513 addi a0,a0,1840 # 21730 <__clzsi2+0x98>
10144: 398000ef jal ra,104dc <printf>
10148: 02c42583 lw a1,44(s0)
1014c: 03442683 lw a3,52(s0)
10150: 03c42783 lw a5,60(s0)
10154: 01c42303 lw t1,28(s0)
10158: 02042883 lw a7,32(s0)
1015c: 02442803 lw a6,36(s0)
10160: 02842503 lw a0,40(s0)
10164: 03042603 lw a2,48(s0)
10168: 03842703 lw a4,56(s0)
1016c: 02b12623 sw a1,44(sp)
10170: 02d12a23 sw a3,52(sp)
10174: 02f12e23 sw a5,60(sp)
10178: 00612e23 sw t1,28(sp)
1017c: 03112023 sw a7,32(sp)
10180: 03012223 sw a6,36(sp)
10184: 02a12423 sw a0,40(sp)
10188: 02c12823 sw a2,48(sp)
1018c: 02e12c23 sw a4,56(sp)
10190: 01c10793 addi a5,sp,28
10194: 03c10693 addi a3,sp,60
10198: 00300593 li a1,3
1019c: 0047a703 lw a4,4(a5)
101a0: 00478793 addi a5,a5,4
101a4: 00e5c5b3 xor a1,a1,a4
101a8: fef69ae3 bne a3,a5,1019c <main+0xd8>
101ac: 00021537 lui a0,0x21
101b0: 74050513 addi a0,a0,1856 # 21740 <__clzsi2+0xa8>
101b4: 328000ef jal ra,104dc <printf>
101b8: 04c12083 lw ra,76(sp)
101bc: 04812403 lw s0,72(sp)
101c0: 00000513 li a0,0
101c4: 05010113 addi sp,sp,80
101c8: 00008067 ret
0001028c <singlenumber>:
1028c: 00100793 li a5,1
10290: 00050693 mv a3,a0
10294: 00052503 lw a0,0(a0)
10298: 02b7d063 bge a5,a1,102b8 <singlenumber+0x2c>
1029c: 00259593 slli a1,a1,0x2
102a0: 00468793 addi a5,a3,4
102a4: 00b686b3 add a3,a3,a1
102a8: 0007a703 lw a4,0(a5)
102ac: 00478793 addi a5,a5,4
102b0: 00e54533 xor a0,a0,a4
102b4: fed79ae3 bne a5,a3,102a8 <singlenumber+0x1c>
102b8: 00008067 ret
```
### Observation
* Line of code: ```80```
* Allocate ```80``` bytes on stack
* Registers used: ```ra``` ```sp``` ```a0 ~ a7``` ```s0 ```
* Number of lw and sw : ```lw: 22``` ```sw: 18 ```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4520
```
## -O3
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 74652 | 2816 |812 |78280 |131c8 |SingleNumber_o3 |
### Header
### Assembly code after compiling
``` clike=
000100c4 <main>:
100c4: 00021537 lui a0,0x21
100c8: ff010113 addi sp,sp,-16
100cc: 00400593 li a1,4
100d0: 66050513 addi a0,a0,1632 # 21660 <__clzsi2+0x88>
100d4: 00112623 sw ra,12(sp)
100d8: 344000ef jal ra,1041c <printf>
100dc: 00021537 lui a0,0x21
100e0: 00c00593 li a1,12
100e4: 67050513 addi a0,a0,1648 # 21670 <__clzsi2+0x98>
100e8: 334000ef jal ra,1041c <printf>
100ec: 00021537 lui a0,0x21
100f0: 00700593 li a1,7
100f4: 68050513 addi a0,a0,1664 # 21680 <__clzsi2+0xa8>
100f8: 324000ef jal ra,1041c <printf>
100fc: 00c12083 lw ra,12(sp)
10100: 00000513 li a0,0
10104: 01010113 addi sp,sp,16
10108: 00008067 ret
000101cc <singlenumber>:
101cc: 00100793 li a5,1
101d0: 00050693 mv a3,a0
101d4: 00052503 lw a0,0(a0)
101d8: 02b7d063 bge a5,a1,101f8 <singlenumber+0x2c>
101dc: 00259593 slli a1,a1,0x2
101e0: 00468793 addi a5,a3,4
101e4: 00b686b3 add a3,a3,a1
101e8: 0007a703 lw a4,0(a5)
101ec: 00478793 addi a5,a5,4
101f0: 00e54533 xor a0,a0,a4
101f4: fef69ae3 bne a3,a5,101e8 <singlenumber+0x1c>
101f8: 00008067 ret
```
### Observation
* Line of code: ```32```
* Allocate ```16``` bytes on stack
* Registers used: ```ra``` ```sp``` ```a0 ~ a5```
* Number of lw and sw : ```lw: 3``` ```sw: 1```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4438
```
## -Ofast
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 74652 | 2816 |812 |78280 |131c8 |SingleNumber_ofast |
### Header
### Assembly code after compiling
``` clike=
000100c4 <main>:
100c4: 00021537 lui a0,0x21
100c8: ff010113 addi sp,sp,-16
100cc: 00400593 li a1,4
100d0: 66050513 addi a0,a0,1632 # 21660 <__clzsi2+0x88>
100d4: 00112623 sw ra,12(sp)
100d8: 344000ef jal ra,1041c <printf>
100dc: 00021537 lui a0,0x21
100e0: 00c00593 li a1,12
100e4: 67050513 addi a0,a0,1648 # 21670 <__clzsi2+0x98>
100e8: 334000ef jal ra,1041c <printf>
100ec: 00021537 lui a0,0x21
100f0: 00700593 li a1,7
100f4: 68050513 addi a0,a0,1664 # 21680 <__clzsi2+0xa8>
100f8: 324000ef jal ra,1041c <printf>
100fc: 00c12083 lw ra,12(sp)
10100: 00000513 li a0,0
10104: 01010113 addi sp,sp,16
10108: 00008067 ret
000101cc <singlenumber>:
101cc: 00100793 li a5,1
101d0: 00050693 mv a3,a0
101d4: 00052503 lw a0,0(a0)
101d8: 02b7d063 bge a5,a1,101f8 <singlenumber+0x2c>
101dc: 00259593 slli a1,a1,0x2
101e0: 00468793 addi a5,a3,4
101e4: 00b686b3 add a3,a3,a1
101e8: 0007a703 lw a4,0(a5)
101ec: 00478793 addi a5,a5,4
101f0: 00e54533 xor a0,a0,a4
101f4: fef69ae3 bne a3,a5,101e8 <singlenumber+0x1c>
101f8: 00008067 ret
```
### Observation
* Line of code: ```32```
* Allocate ```16``` bytes on stack
* Registers used: ```ra``` ```sp``` ```a0 ~ a5```
* Number of lw and sw : ```lw: 3``` ```sw: 1```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4438
```
## -Os
### Size
| text | data | bss |dec |hex |filename |
| -------- | -------- | -------- | -------- | -------- | -------- |
| 74836 | 2816 |812 |78464 |13280 |SingleNumber_os |
### Header
### Assembly code after compiling
``` clike=
000100c4 <main>:
100c4: f9010113 addi sp,sp,-112
100c8: 000215b7 lui a1,0x21
100cc: 06812423 sw s0,104(sp)
100d0: 01400613 li a2,20
100d4: 6f058413 addi s0,a1,1776 # 216f0 <__clzsi2+0xb8>
100d8: 00c10513 addi a0,sp,12
100dc: 6f058593 addi a1,a1,1776
100e0: 06112623 sw ra,108(sp)
100e4: 27c000ef jal ra,10360 <memcpy>
100e8: 00500593 li a1,5
100ec: 00c10513 addi a0,sp,12
100f0: 140000ef jal ra,10230 <singlenumber>
100f4: 00050593 mv a1,a0
100f8: 00021537 lui a0,0x21
100fc: 6c050513 addi a0,a0,1728 # 216c0 <__clzsi2+0x88>
10100: 520000ef jal ra,10620 <printf>
10104: 01c00613 li a2,28
10108: 01440593 addi a1,s0,20
1010c: 02010513 addi a0,sp,32
10110: 250000ef jal ra,10360 <memcpy>
10114: 00700593 li a1,7
10118: 02010513 addi a0,sp,32
1011c: 114000ef jal ra,10230 <singlenumber>
10120: 00050593 mv a1,a0
10124: 00021537 lui a0,0x21
10128: 6d050513 addi a0,a0,1744 # 216d0 <__clzsi2+0x98>
1012c: 4f4000ef jal ra,10620 <printf>
10130: 02400613 li a2,36
10134: 03040593 addi a1,s0,48
10138: 03c10513 addi a0,sp,60
1013c: 224000ef jal ra,10360 <memcpy>
10140: 00900593 li a1,9
10144: 03c10513 addi a0,sp,60
10148: 0e8000ef jal ra,10230 <singlenumber>
1014c: 00050593 mv a1,a0
10150: 00021537 lui a0,0x21
10154: 6e050513 addi a0,a0,1760 # 216e0 <__clzsi2+0xa8>
10158: 4c8000ef jal ra,10620 <printf>
1015c: 06c12083 lw ra,108(sp)
10160: 06812403 lw s0,104(sp)
10164: 00000513 li a0,0
10168: 07010113 addi sp,sp,112
1016c: 00008067 ret
00010230 <singlenumber>:
10230: 00050713 mv a4,a0
10234: 00052503 lw a0,0(a0)
10238: 00100793 li a5,1
1023c: 00b7c463 blt a5,a1,10244 <singlenumber+0x14>
10240: 00008067 ret
10244: 00279693 slli a3,a5,0x2
10248: 00d706b3 add a3,a4,a3
1024c: 0006a683 lw a3,0(a3)
10250: 00178793 addi a5,a5,1
10254: 00d54533 xor a0,a0,a3
10258: fe5ff06f j 1023c <singlenumber+0xc>
```
### Observation
* Line of code: ```56```
* Allocate ```112``` bytes on stack
* Registers used: ```ra``` ```sp``` ```a0 ~ a5``` ```s0```
* Number of lw and sw : ```lw: 4``` ```sw: 2```
### CSR cycle count and Execution Result
```
result 1: 4
result 2: 5
result 3: 7
inferior exit code 0
CSR cycle count: 4675
```
## Summary
### Summary of results
| | line of code | Allocate bytes on stack | Registers used |lw|sw|CSR|
| -------- | -------- | -------- | -------- | -------- | -------- | -------- |
| O0 |126 | 144 |```ra``` ```sp``` ```a0 ~ a7``` ```s0``` ```t1``` |40 |38 |4792|
| O1 | 126 | 144 |```ra``` ```sp``` ```a0 ~ a7``` ```s0``` ```t1``` |40 |38 |4562 |
| O2 | 80 | 80 |```ra``` ```sp``` ```a0 ~ a7``` ```s0``` |22 |18 |4520 |
| O3 | 32 | 16 |```ra``` ```sp``` ```a0 ~ a7``` |3 |1 |4438 |
| Ofast | 32 | 16 |```ra``` ```sp``` ```a0 ~ a7``` |3 |1 |4438 |
| Os | 56 | 112 |```ra``` ```sp``` ```a0 ~ a7``` ``s0`` |4 |2 |4675 |
### anaylize
* **O0** to **O1** : Almost the same, but fewer CSR.
* **O1** to **O2** : Fewer lines of Code, stack usage amount, lw/sw count, CSR. The register ```t1```not used since O2.
* **O2** to **O3** : More Fewer lines of Code, stack usage amount, lw/sw count, CSR. The register ```s0``` is not used in O3. Futhermore, O3's main funtion address is in front of the singlenumber fuction.
* **Ofast** :As same as O3, not any differance.
* **Os**: Lines of Code, stack usage amount, lw/sw count, register usage amount.
is between O2 and O3, but CSR is between O0 and O1.
Google Drive
Gist
Clipboard
Sign in with Wallet
