Float Product combine1: Maximum use of data abstraction:
10.51 cycles/element
Float Product combine2: Take vec_length() out of loop:
15.68 cycles/element
Float Product combine3: Array reference to vector data:
6.70 cycles/element
Float Product combine4: Array reference, accumulate in temporary:
6.67 cycles/element
Float Product combine4p: Pointer reference, accumulate in temporary:
6.67 cycles/element
Float Product Array code, unrolled by 2:
6.67 cycles/element
Float Product combine5p: Pointer code, unrolled by 3, for loop:
6.66 cycles/element
Float Product Array code, unrolled by 3, while loop:
6.66 cycles/element
Float Product Array code, unrolled by 4:
6.67 cycles/element
Float Product Array code, unrolled by 8:
6.68 cycles/element
Float Product Array code, unrolled by 16:
6.67 cycles/element
Float Product Pointer code, unrolled by 2:
6.67 cycles/element
Float Product Pointer code, unrolled by 3:
6.67 cycles/element
Float Product Pointer code, unrolled by 4:
6.67 cycles/element
Float Product Pointer code, unrolled by 8:
6.68 cycles/element
Float Product Pointer code, unrolled by 16:
6.67 cycles/element
Float Product combine6: Array code, unrolled by 2, Superscalar x2:
3.34 cycles/element
Float Product Array code, unrolled by 4, Superscalar x2:
3.33 cycles/element
Float Product Array code, unrolled by 8, Superscalar x2:
3.37 cycles/element
Float Product Array code, unrolled by 3, Superscalar x3:
2.23 cycles/element
Float Product Array code, unrolled by 4, Superscalar x4:
1.67 cycles/element
Float Product Array code, unrolled by 8, Superscalar x4:
1.74 cycles/element
Float Product Array code, unrolled by 6, Superscalar x6:
1.71 cycles/element
Float Product Array code, unrolled by 8, Superscalar x8:
1.75 cycles/element
Float Product Array code, unrolled by 10, Superscalar x10:
1.74 cycles/element
Float Product Array code, unrolled by 12, Superscalar x6:
1.74 cycles/element
Float Product Array code, unrolled by 12, Superscalar x12:
1.73 cycles/element
Float Product Pointer code, unrolled by 8, Superscalar x2:
3.43 cycles/element
Float Product Pointer code, unrolled by 8, Superscalar x4:
1.73 cycles/element
Float Product Pointer code, unrolled by 8, Superscalar x8:
1.73 cycles/element
Float Product Pointer code, unrolled by 9, Superscalar x3:
2.29 cycles/element
Float Product Array code, Unroll x2, Superscalar x2, noninterleaved:
3.34 cycles/element
Float Product Array code, unrolled by 2, different associativity:
3.33 cycles/element
Float Product Array code, unrolled by 3, Different Associativity:
2.23 cycles/element
Float Product Array code, unrolled by 4, Different Associativity:
1.68 cycles/element
Float Product Array code, unrolled by 6, Different Associativity:
1.71 cycles/element
Float Product Array code, unrolled by 8, Different Associativity:
1.72 cycles/element
Float Product SSE code, 1*VSIZE-way parallelism:
1.68 cycles/element
Float Product SSE code, 2*VSIZE-way parallelism:
0.94 cycles/element
Float Product SSE code, 4*VSIZE-way parallelism:
0.40 cycles/element
Float Product SSE code, 8*VSIZE-way parallelism:
0.41 cycles/element
Float Product SSE code, 12*VSIZE-way parallelism:
0.39 cycles/element
Float Product SSE code, 2*VSIZE-way parallelism, reassociate:
0.86 cycles/element
Float Product SSE code, 4*VSIZE-way parallelism, reassociate:
0.41 cycles/element
Float Product SSE code, 8*VSIZE-way parallelism, reassociate:
0.42 cycles/element