---

D3D9
[RwMatrix]

Platform specific

RwMatrixMultiply

Note:

This section only applies if the RenderWare SDK has been built with available assembler code. This is the default case.

The RwMatrixMultiply function has been changed in D3D9 RenderWare Graphics to utilize x96 floating-point assembler code to calculate the product of two RwMatrix's. This assembler is written to take advantage of the structure of an RwMatrix and examples requiring a large number of matrix products will see a performance improvement. For example, in applications that use RpSkin.

Note:

This assembler code does not take advantage of any vector instructions available on the host processor. It is, however, written to use instruction pairing/overlapping as available on Pentium (and equivalent) processors.

---