![]() |
AI Engine Intrinsics User Guide
(AIE) v(2024.1)
|
Float vector comparison.
Performs the comparison between lanes and return the result of the comparison as a bit in the return word.
Functions | |
| unsigned int | fpge (v8float acc, v8float xbuf, int xstart, unsigned int xoffs) |
| Greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fpge (v8float acc, v16float xbuf, int xstart, unsigned int xoffs) |
| Greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fpge (v8float acc, v32float xbuf, int xstart, unsigned int xoffs) |
| Greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fpge_abs (v8float acc, v8float xbuf, int xstart, unsigned int xoffs) |
| Take absolute value and perform greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fpge_abs (v8float acc, v8float xbuf) |
| Take absolute value and perform greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fpge_abs (v8float acc, v32float xbuf, int xstart, unsigned int xoffs) |
| Take absolute value and perform greater-equal comparison for single precision real floating point vectors. | |
| unsigned int | fplt (v8float acc, v8float xbuf, int xstart, unsigned int xoffs) |
| Less-than comparision for single precision real floating point vectors. | |
| unsigned int | fplt (v8float acc, v16float xbuf, int xstart, unsigned int xoffs) |
| Less-than comparision for single precision real floating point vectors. | |
| unsigned int | fplt (v8float acc, v32float xbuf, int xstart, unsigned int xoffs) |
| Less-than comparision for single precision real floating point vectors. | |
| unsigned int | fplt_abs (v8float acc, v8float xbuf, int xstart, unsigned int xoffs) |
| Take absolute value and perform less-than comparison for single precision real floating point vectors. | |
| unsigned int | fplt_abs (v8float acc, v8float xbuf) |
| Take absolute value and perform less-than comparison for single precision real floating point vectors. | |
| unsigned int | fplt_abs (v8float acc, v32float xbuf, int xstart, unsigned int xoffs) |
| Take absolute value and perform less-than comparison for single precision real floating point vectors. | |
Variables | |
| return | cmp |
| v8float | ret |
| v16float | xbuf |
| v16float int unsigned int | xoffs |
| v16float int | xstart |
| unsigned int fpge | ( | v8float | acc, |
| v8float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] >= xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fpge | ( | v8float | acc, |
| v16float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] >= xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fpge | ( | v8float | acc, |
| v32float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] >= xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fpge_abs | ( | v8float | acc, |
| v8float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Take absolute value and perform greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] >= abs(xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fpge_abs | ( | v8float | acc, |
| v8float | xbuf | ||
| ) |
Take absolute value and perform greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0 ; i < 8 ; i++) ret[i] = acc[i] >= abs(xbuf[i] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| unsigned int fpge_abs | ( | v8float | acc, |
| v32float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Take absolute value and perform greater-equal comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] >= abs(xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fplt | ( | v8float | acc, |
| v8float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Less-than comparision for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] < xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fplt | ( | v8float | acc, |
| v16float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Less-than comparision for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] < xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fplt | ( | v8float | acc, |
| v32float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Less-than comparision for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] < xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fplt_abs | ( | v8float | acc, |
| v8float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Take absolute value and perform less-than comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] < abs(xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| unsigned int fplt_abs | ( | v8float | acc, |
| v8float | xbuf | ||
| ) |
Take absolute value and perform less-than comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0 ; i < 8 ; i++) ret[i] = acc[i] < abs(xbuf[i] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| unsigned int fplt_abs | ( | v8float | acc, |
| v32float | xbuf, | ||
| int | xstart, | ||
| unsigned int | xoffs | ||
| ) |
Take absolute value and perform less-than comparison for single precision real floating point vectors.
~~~~~~~~~~~~~~~~~~~ for (i = 0; i < 8; i++) ret[i] = acc[i] < abs(xbuf[xstart + xoffs[i]] ~~~~~~~~~~~~~~~~~~~
| acc | Incoming accumulation vector. |
| xbuf | Input buffer. |
| xstart | Starting offset for all lanes of X. |
| xoffs | 8 x 4 bits: Additional lane-dependent offset for X. |
| return cmp |
| v8float ret |
| v16float xbuf |
| v16float int unsigned int xoffs |
| v16float int xstart |