 |
AI Engine Intrinsics User Guide (AIE) v2024.2
|
|
AI Engine Intrinsics User Guide (AIE) v2024.2
|
Simple intrinsics for buffer self-multiplication. For more information in lane selection please refer to here.
Functions | |
| v8acc48 | mac8 (v8acc48 acc, v8int16 xbuff) |
| Self multiplication with accumulate on a single buffer (equivalent to lane by lane power of 2). For more information in lane selection please refer to here. | |
| v8acc48 | mul8 (v8int16 xbuff) |
| Self multiplication on a single buffer (equivalent to lane by lane power of 2). For more information in lane selection please refer to here. | |
Without squares | |
These intrinsics have a default x/y square of 0x3210. | |
| v8acc48 | mac8 (v8acc48 acc, v64int16 xbuff, int xstart, int xoffsets, int xstep, int zstart, int zoffsets, int zstep) |
| Self multiplication with accumulate on a single buffer. For more information in lane selection please refer to here. | |
| v8acc48 | mac8 (v8acc48 acc, v32int16 xbuff0, v32int16 xbuff1, int xstart, int xoffsets, int xstep, int zstart, int zoffsets, int zstep) |
| Self multiplication with accumulate combining two buffers into one. For more information in lane selection please refer to here. | |
| v8acc48 | mul8 (v64int16 xbuff, int xstart, int xoffsets, int xstep, int zstart, int zoffsets, int zstep) |
| Self multiplication on a single buffer. For more information in lane selection please refer to here. | |
| v8acc48 mac8 | ( | v8acc48 | acc, |
| v32int16 | xbuff0, | ||
| v32int16 | xbuff1, | ||
| int | xstart, | ||
| int | xoffsets, | ||
| int | xstep, | ||
| int | zstart, | ||
| int | zoffsets, | ||
| int | zstep | ||
| ) |
Self multiplication with accumulate combining two buffers into one. For more information in lane selection please refer to here.
| v8acc48 mac8 | ( | v8acc48 | acc, |
| v64int16 | xbuff, | ||
| int | xstart, | ||
| int | xoffsets, | ||
| int | xstep, | ||
| int | zstart, | ||
| int | zoffsets, | ||
| int | zstep | ||
| ) |
Self multiplication with accumulate on a single buffer. For more information in lane selection please refer to here.
| v8acc48 mac8 | ( | v8acc48 | acc, |
| v8int16 | xbuff | ||
| ) |
Self multiplication with accumulate on a single buffer (equivalent to lane by lane power of 2). For more information in lane selection please refer to here.
| v8acc48 mul8 | ( | v64int16 | xbuff, |
| int | xstart, | ||
| int | xoffsets, | ||
| int | xstep, | ||
| int | zstart, | ||
| int | zoffsets, | ||
| int | zstep | ||
| ) |
Self multiplication on a single buffer. For more information in lane selection please refer to here.
| v8acc48 mul8 | ( | v8int16 | xbuff | ) |
Self multiplication on a single buffer (equivalent to lane by lane power of 2). For more information in lane selection please refer to here.