8 | | * [890] Do the Taylor series on `x / π` instead of `x`: results in a measurable speed increase in `sin()` (2% with ppu-lv2-g++, almost 1% with SNC) |
9 | | * [891] Do the Taylor series in two steps, for better pipelining at the cost of one extra multiplication: a notable speed increase in `sin()` (11% with ppu-lv2-g++, 2% with SNC) |
| 8 | * [890] Do the Taylor series on `x / π` instead of `x`: results in a measurable speed increase in `sin()` |
| 9 | * 2% with ppu-lv2-g++ 4.1.1 on a PS3 |
| 10 | * almost 1% with SNC 360.1 on a PS3 |
| 11 | * 30% with g++ 4.5.2 on mingw32 running on a Core2Quad |
| 12 | * [891] Do the Taylor series in two steps, for better pipelining at the cost of one extra multiplication: a notable speed increase in `sin()` |
| 13 | * 11% with ppu-lv2-g++ 4.1.1 on a PS3 |
| 14 | * 2% with SNC 360.1 on a PS3 |
| 15 | * 7% with g++ 4.5.2 on mingw32 running on a Core2Quad |