Performance of VML Functions
| Default processor | Intel® Pentium® III processor | Intel® Pentium® 4 processor | Intel® Itanium® 2 processor | |||||||||||||
| float | double | float | double | float | double | float | double | |||||||||
| HA | LA | HA | LA | HA | LA | HA | LA | HA | LA | HA | LA | HA | LA | HA | LA | |
| Inv | 9.90 | 9.90 | 11.24 | 11.24 | 10.12 | 10.12 | 11.31 | 11.23 | 9.85 | 3.47 | 11.75 | 10.17 | 3.11 | 3.11 | 4.11 | 4.11 |
| Div | 14.59 | 14.59 | 16.27 | 16.24 | 11.07 | 10.96 | 16.43 | 16.33 | 9.83 | 3.74 | 11.68 | 11.68 | 4.11 | 4.11 | 5.18 | 5.18 |
| Sqrt | 20.61 | 20.61 | 34.83 | 34.83 | 15.73 | 15.73 | 33.14 | 33.05 | 9.41 | 4.94 | 27.08 | 25.71 | 5.12 | 5.12 | 7.15 | 7.15 |
| InvSqrt | 21.13 | 21.07 | 32.25 | 32.25 | 8.16 | 8.13 | 31.60 | 31.50 | 6.20 | 3.75 | 19.61 | 15.80 | 5.12 | 5.12 | 6.13 | 6.13 |
| Cbrt | 30.75 | 30.52 | 45.22 | 45.22 | 30.05 | 30.05 | 45.34 | 45.34 | 22.42 | 17.29 | 40.67 | 32.42 | 7.16 | 7.16 | 10.20 | 10.20 |
| InvCbrt | 30.85 | 30.85 | 42.33 | 42.15 | 31.30 | 31.25 | 42.31 | 42.03 | 23.43 | 15.83 | 40.72 | 26.87 | 7.17 | 7.17 | 9.18 | 9.18 |
| Pow | 108.54 | 68.20 | 144.46 | 96.41 | 106.83 | 66.51 | 145.14 | 95.92 | 38.82 | 38.82 | 105.93 | 79.24 | 11.26 | 11.26 | 21.36 | 21.36 |
| Powx | 106.25 | 106.25 | 144.22 | 144.22 | 105.98 | 105.98 | 143.88 | 143.88 | 39.22 | 39.22 | 81.36 | 80.48 | 9.44 | 9.44 | 21.44 | 21.43 |
| Exp | 23.02 | 22.76 | 36.92 | 36.92 | 19.93 | 19.93 | 37.10 | 37.07 | 10.76 | 9.22 | 26.14 | 16.74 | 4.16 | 4.15 | 6.17 | 6.17 |
| Ln | 20.79 | 20.79 | 35.25 | 35.25 | 20.83 | 20.83 | 37.72 | 37.60 | 15.78 | 12.58 | 25.18 | 23.91 | 7.17 | 7.17 | 11.19 | 11.19 |
| Log10 | 20.72 | 20.72 | 35.26 | 35.19 | 20.97 | 20.86 | 37.65 | 37.58 | 19.07 | 12.82 | 25.68 | 24.57 | 7.17 | 7.17 | 11.20 | 11.20 |
| Cos | 44.12 | 44.12 | 52.16 | 52.16 | 44.33 | 30.48 | 52.28 | 52.28 | 22.36 | 12.07 | 35.89 | 35.87 | 7.17 | 7.17 | 9.19 | 9.19 |
| Sin | 40.64 | 40.64 | 49.71 | 49.64 | 42.12 | 26.70 | 48.90 | 48.36 | 19.51 | 11.14 | 35.94 | 35.82 | 6.15 | 6.15 | 8.17 | 8.17 |
| SinCos | 64.74 | 64.48 | 77.67 | 77.67 | 65.07 | 39.80 | 77.05 | 77.05 | 29.39 | 19.26 | 64.10 | 48.12 | 8.18 | 8.18 | 11.20 | 11.20 |
| Tan | 58.90 | 58.90 | 75.28 | 68.69 | 59.66 | 38.97 | 75.60 | 68.18 | 40.35 | 18.73 | 66.48 | 47.92 | 9.21 | 9.21 | 11.24 | 11.24 |
| Acos | 70.42 | 70.30 | 98.41 | 97.47 | 43.16 | 36.17 | 98.28 | 97.36 | 25.06 | 16.17 | 69.17 | 51.66 | 10.22 | 10.22 | 15.30 | 15.30 |
| Asin | 66.27 | 66.00 | 93.95 | 93.95 | 39.71 | 28.30 | 94.11 | 94.11 | 23.47 | 15.92 | 65.42 | 65.42 | 10.21 | 10.21 | 15.29 | 15.29 |
| Atan | 65.65 | 65.65 | 87.78 | 87.65 | 37.72 | 19.74 | 87.34 | 87.16 | 24.48 | 14.43 | 62.51 | 50.50 | 11.39 | 11.39 | 13.31 | 13.31 |
| Atan2 | 85.59 | 85.59 | 139.92 | 139.85 | 78.98 | 32.94 | 139.96 | 139.96 | 56.10 | 28.60 | 113.27 | 67.22 | 12.32 | 12.32 | 19.45 | 19.45 |
| Cosh | 49.83 | 49.80 | 72.17 | 72.17 | 37.97 | 37.97 | 72.36 | 72.36 | 20.48 | 12.98 | 37.34 | 27.39 | 7.17 | 7.17 | 10.20 | 10.20 |
| Sinh | 50.91 | 50.91 | 71.02 | 71.00 | 43.53 | 43.53 | 71.32 | 71.32 | 32.71 | 14.65 | 39.40 | 29.26 | 7.16 | 7.16 | 12.19 | 12.19 |
| Tanh | 52.22 | 52.21 | 107.74 | 107.73 | 52.23 | 52.23 | 107.77 | 107.75 | 22.73 | 21.19 | 56.31 | 39.69 | 6.62 | 6.62 | 10.34 | 10.34 |
| Acosh | 153.27 | 153.27 | 111.04 | 111.04 | 43.68 | 43.68 | 110.99 | 110.99 | 31.08 | 22.53 | 88.92 | 64.11 | 13.25 | 13.25 | 18.32 | 18.32 |
| Asinh | 164.95 | 163.92 | 115.94 | 115.94 | 46.91 | 45.83 | 117.23 | 117.23 | 32.92 | 21.42 | 104.30 | 79.52 | 14.31 | 14.31 | 18.40 | 18.40 |
| Atanh | 110.00 | 110.00 | 96.83 | 96.65 | 36.39 | 36.39 | 97.93 | 97.93 | 39.03 | 22.54 | 92.03 | 73.89 | 12.28 | 12.28 | 19.34 | 19.34 |
| Erf | 54.74 | 54.74 | 110.21 | 109.63 | 55.50 | 55.50 | 110.78 | 110.25 | 24.30 | 21.01 | 55.72 | 39.44 | 6.37 | 6.34 | 10.34 | 10.33 |
| Erfc | 258.04 | 257.85 | 257.73 | 254.85 | 248.00 | 248.00 | 258.13 | 254.77 | 45.14 | 39.42 | 96.68 | 66.97 | 8.20 | 8.20 | 16.22 | 16.22 |
Notes:
1) Units - CPE (Clocks per Element)
2) Data - vectors of 1000 elements with random generated numbers
3) "Default" means X87 code for all IA-32 processors
4) Performance of "default" version was measured on
Pentium® III processor
Copyright © 2000-2003, Intel Corporation, All Rights Reserved.