Integer Sum combine1: Maximum use of data abstraction:
15.00 cycles/element
Integer Sum combine2: Take vec_length() out of loop:
10.00 cycles/element
Integer Sum combine3: Array reference to vector data:
6.29 cycles/element
Integer Sum combine4: Array reference, accumulate in temporary:
2.33 cycles/element
Integer Sum combine4p: Pointer reference, accumulate in temporary:
3.00 cycles/element
Integer Sum Array code, unrolled by 2:
1.50 cycles/element
Integer Sum combine5p: Pointer code, unrolled by 3, for loop:
1.01 cycles/element
Integer Sum Array code, unrolled by 3, while loop:
1.00 cycles/element
Integer Sum Array code, unrolled by 4:
0.75 cycles/element
Integer Sum Array code, unrolled by 8:
0.62 cycles/element
Integer Sum Array code, unrolled by 16:
0.81 cycles/element
Integer Sum Pointer code, unrolled by 2:
1.00 cycles/element
Integer Sum Pointer code, unrolled by 3:
1.00 cycles/element
Integer Sum Pointer code, unrolled by 4:
1.00 cycles/element
Integer Sum Pointer code, unrolled by 8:
1.00 cycles/element
Integer Sum Pointer code, unrolled by 16:
0.99 cycles/element
Integer Sum combine6: Array code, unrolled by 2, Superscalar x2:
1.50 cycles/element
Integer Sum Array code, unrolled by 4, Superscalar x2:
0.83 cycles/element
Integer Sum Array code, unrolled by 8, Superscalar x2:
0.63 cycles/element
Integer Sum Array code, unrolled by 3, Superscalar x3:
1.00 cycles/element
Integer Sum Array code, unrolled by 4, Superscalar x4:
1.01 cycles/element
Integer Sum Array code, unrolled by 8, Superscalar x4:
0.63 cycles/element
Integer Sum Array code, unrolled by 6, Superscalar x6:
0.83 cycles/element
Integer Sum Array code, unrolled by 8, Superscalar x8:
0.63 cycles/element
Integer Sum Array code, unrolled by 10, Superscalar x10:
0.60 cycles/element
Integer Sum Array code, unrolled by 12, Superscalar x6:
0.67 cycles/element
Integer Sum Array code, unrolled by 12, Superscalar x12:
0.86 cycles/element
Integer Sum Pointer code, unrolled by 8, Superscalar x2:
0.63 cycles/element
Integer Sum Pointer code, unrolled by 8, Superscalar x4:
0.69 cycles/element
Integer Sum Pointer code, unrolled by 8, Superscalar x8:
0.69 cycles/element
Integer Sum Pointer code, unrolled by 9, Superscalar x3:
0.56 cycles/element
Integer Sum Array code, Unroll x2, Superscalar x2, noninterleaved:
1.11 cycles/element
Integer Sum Array code, unrolled by 2, different associativity:
1.50 cycles/element
Integer Sum Array code, unrolled by 3, Different Associativity:
1.00 cycles/element
Integer Sum Array code, unrolled by 4, Different Associativity:
1.00 cycles/element
Integer Sum Array code, unrolled by 6, Different Associativity:
0.83 cycles/element
Integer Sum Array code, unrolled by 8, Different Associativity:
0.81 cycles/element
Integer Sum SSE code, 1*VSIZE-way parallelism:
0.75 cycles/element
Integer Sum SSE code, 2*VSIZE-way parallelism:
0.38 cycles/element
Integer Sum SSE code, 4*VSIZE-way parallelism:
0.28 cycles/element
Integer Sum SSE code, 8*VSIZE-way parallelism:
0.27 cycles/element
Integer Sum SSE code, 12*VSIZE-way parallelism:
0.26 cycles/element
Integer Sum SSE code, 2*VSIZE-way parallelism, reassociate:
0.42 cycles/element
Integer Sum SSE code, 4*VSIZE-way parallelism, reassociate:
0.30 cycles/element
Integer Sum SSE code, 8*VSIZE-way parallelism, reassociate:
0.30 cycles/element