|
Operation |
PPro 200 |
PII 300 |
| Vector dot product, 1000 elements |
1.06 sec (19 Mflops) |
0.080 sec (250 Mflops) |
|
Matrix-vector multiply, 50 elements |
5.35 sec (19 Mflops) |
0.90 sec (111 Mflops) |
|
Matrix multiply, 100x100, unrolled |
4.55 sec (22 Mflops) |
1.08 sec (93 Mflops) |
|
Trigonometric function calls |
2.76 sec -- |
1.84 sec -- |
|
Scalar times vector, 1000 elements |
3.16 sec (32 Mflops) |
0.70 sec (143 Mflops) |
|
Scalar times vector, 10 elements |
3.55 sec (28 Mflops) |
0.46 sec (217 Mflops) |
Compiled with Absoft native f77 compiler (the Absoft compiler is ~35% faster than gnu g77).
All benchmarks run in double precision.