We use cpu (and fpu) optimalizations for mathematical calculations like matrixes.
AMD 3d now is supported parallel by the SSE instructions. It could be strong in matrix calculations like multiply.
The DDDengine does not require the extended cpu, so it can skip the optimalizations, but all tool can run in SSE2 supported cpus only.
The most optimalization not require the SSE2, so the SSE is still enought to enjoy this extensions, but the SSE3 has some new options to make more power on matrix calculations, by the way our builder does not supported it yet, sorry...
Feel the power of x64 (x64-86).
There are 3 layers for the applications:
- Driver: all driver must be in 64bits on x64.
Some new driver make more speed on x64.
- Interface to drivers: windows API, directX, and OpenGL
The new DirectX 10 speed up 10-15% on x64.
- Our techonlogy add more than 10% on x64 also.
How fast the x64 system?
The x64 system has a different code than the x86, but we need optimize our programs to use ths extra performance.
- the x64 expand the memory limt over the 4 GBytes.
- the memory will be twice wider under x64.
- new registers help to optimalize our code.
- the exdended mathematic could be simply on x64, like MMX, SSE, and double arithmetic.