https://www.geeksforgeeks.org/structure-member-alignment-padding-and-data-packing/

参考资料

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

module OneHotDecoder #(
    parameter integer NUM_BITS = 16
)(
    input[NUM_BITS-1:0]             bits,
    output[$clog2(NUM_BITS)-1:0]    idx
);

wire [$clog2(NUM_BITS)-1:0] stage[0:NUM_BITS];
assign stage[0] = 0;// desired default output if not bits set

genvar i;
generate
for (i = 0; i < NUM_BITS; i = i+1) begin
    assign stage[i+1] = bits[i +: 1] ? i : stage[i];
end
endgenerate

assign idx = stage[NUM_BITS];

endmodule // OneHotDecoder

References

• https://stackoverflow.com/a/38258865/5416913

Creating Example Project

Custom AXI Slave Module

Custom AXI-Lite Slave Module

Custom AXI VIP

AES Decryption

AES decryption is the reverse version of encryption. Here “reverse” has three meanings.

1. Reverse means round keys are used in reverse order, i.e. RoundKey[10] is first used, the RoundKey[0] is last used. Encryption and decryption use the same round keys.
2. Reverse means the order of the four steps in each encryption round is reverse.
3. Reverse means row shifting orientation is reverse in Shift Rows step.

Introduction

Vivado HLS has been widely used due to its feasibility, however, how to write a fully optimized HLS is still a challenge. I will use matrix multiplication as an example to illustrate how to optimize HLS code.

This blog has two goals:

1. I would like to summarize a methodology about how to optimize HLS code step by step.
2. I would like to use this blog as a summary to summarize common matrix multiplication codes.