AI Engine Intrinsics
(AIE) r2p21
|
Vector MAC operations allowing pre-subtraction from within the X buffer, or between the X and Y buffers. The data for the final column for each lane can be selected seperately and without pre-subtraction for additional flexibility. The result is then multiplied with data from the Z buffer.
For general information on all vector MAC intrinsics go here.
Modules | |
16 bit x 16 bit | |
16 bit x 32 bit | |
32 bit x 16 bit | |
32 bit x 32 bit | |