CODEWITHSUNDEEP
cassemblyriscvinstruction-set
exiturnor
exiturnor

Reputation: 57

Use of SLLI for accessing array elements in loop

This is the example my teacher gave in C:

while (save[i] == k) i+=1;

Compiled in RISC-V mode, where i is in x22, k in x24, save address is in x25:

slli x10,x22,3                
add  x10,x10,x25   
ld   x9, 0(x10)   
bne  x9,x24, EXIT  
addi x22,x22,1  
beq  x0,x0, Loop  
Exit:.....

I don't understand why he used the shift left by immediate (slli)

Upvotes: 3

Views: 4232

Answers (1)

jfMR
jfMR

Reputation: 24738

I don't understand why he used the shift left by immediate (slli)

Left shifting a number N bits produces the same result as multiplying the number by 2N. Since a multiply instruction is generally more expensive (slower) than a shift instruction, left shifting is used instead of a multiply instruction as an optimization when possible: if one of the operands is a constant power of two.

slli is used to calculate the offset into the array save. Since i corresponds to x22, then:

slli x10,x22,3

multiplies the value of i by 8 (i.e., 23, a power of two) and writes this result to x10.

x25 indicates the base address of the array save – the address of its first element – save[0]. Together with x10, which corresponds to the offset, is used to calculate the address of save[i]:

add  x10,x10,x25

At this point, x10 holds the address of save[i]. That element is eventually loaded into x9:

ld   x9, 0(x10)

Upvotes: 3

Related Questions