Vector Input - 2024.2 English - XD160

Vitis Libraries
Document ID
XD160
Release Date
2024-11-29
Version
2024.2 English

The design of the primitive includes three modules:

  1. read: Reads data from the input stream then output data by one stream whose width is lcm(Win, N * Wout) bits. Here, the least common multiple of Win and N * Wout is the inner buffer size to solve the different input width and output width.
  2. reduce: Splits the large width to an array of N elements of Wout bits.
  3. distribute: Reads the array of elements, and distributes them to output streams that are not yet full.
design details of n streams to one distribution on load balance

Attention

Current implementation has the following limitations:

  • It uses a wide ap_uint as an internal buffer. The buffer is as wide as the least common multiple (LCM) of input width and total output width. The width is limited by AP_INT_MAX_W, which defaults to 1024.
  • This library tries to override AP_INT_MAX_W to 4096. Ensure that ap_int.h is not included before the library headers.
  • Too large AP_INT_MAX_W significantly slows down HLS synthesis.

Important

The depth of output streams must be no less than four due to an internal delay.