1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
|
//
// Copyright 2011-2014 Ettus Research LLC
// Copyright 2018 Ettus Research, a National Instruments Company
//
// SPDX-License-Identifier: LGPL-3.0-or-later
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
//
module hb47_int
#(parameter WIDTH=18,
parameter DEVICE="7SERIES")
(input clk,
input rst,
input bypass,
input stb_in,
input [WIDTH-1:0] data_in,
input [7:0] output_rate,
input stb_out,
output reg [WIDTH-1:0] data_out);
// Input data Pipeline
reg [WIDTH-1:0] data_in_pipe[0:23];
reg stb_pipe0, stb_pipe1, stb_pipe2, stb_pipe3, stb_pipe4 ;
reg stb_pipe5, stb_pipe6, stb_pipe7, stb_pipe8, stb_pipe9 ;
wire [WIDTH-1:0] sample_a[0:5], sample_b[0:5];
wire [17:0] coeff[0:5];
wire [47:0] accumulator_out[0:5];
reg [47:0] partial_result_a, partial_result_b;
reg [47:0] result;
// Implicit coeff23 of 131071
wire [17:0] coeff_a[0:5];
wire [17:0] coeff_b[0:5];
assign coeff_a[0] = -62;
assign coeff_b[0] = 194;
assign coeff_a[1] = -440;
assign coeff_b[1] = 855;
assign coeff_a[2] = -1505;
assign coeff_b[2] = 2478;
assign coeff_a[3] = -3900;
assign coeff_b[3] = 5990;
assign coeff_a[4] = -9187;
assign coeff_b[4] = 14632;
assign coeff_a[5] = -26536;
assign coeff_b[5] = 83009;
genvar i;
always @(posedge clk)
if (rst)
data_in_pipe[0] <= 18'h0;
else if (stb_in)
data_in_pipe[0] <= data_in;
generate
for (i=0; i<23; i=i+1) begin: sample_pipeline
always @(posedge clk) if (rst)
data_in_pipe[i+1] <= 0;
else if (stb_in)
data_in_pipe[i+1] <= data_in_pipe[i];
end
endgenerate
generate
for (i=0; i<6; i=i+1) begin: filter_core
assign sample_a[i] = stb_pipe0 ? data_in_pipe[(i*2)] : data_in_pipe[(2*i)+1];
assign sample_b[i] = stb_pipe0 ? data_in_pipe[23-(i*2)] : data_in_pipe[23-(2*i)-1];
// Coeffs are 1 pipeline downstream of sample input
assign coeff[i] = stb_pipe1 ? coeff_a[i] : coeff_b[i];
add_then_mac
#(.DEVICE(DEVICE))
add_then_mac_i
(
.acc(accumulator_out[i]),
.carryin(1'b0),
.ce(1'b1),
.clk(clk),
.b(coeff[i]),
.load(stb_pipe2),
.c(48'h0),
.a(sample_a[i]),
.d(sample_b[i]),
.rst(rst)
);
end // block: filter_core
endgenerate
//
// Dual 3:1 compressors
//
always @(posedge clk) if (stb_pipe5) begin
partial_result_a[47:0] <= accumulator_out[0] + accumulator_out[1] + accumulator_out[2];
partial_result_b[47:0] <= accumulator_out[3] + accumulator_out[4] + accumulator_out[5];
end
//
// Final Result Adder
//
always @(posedge clk) if (stb_pipe6) begin
result[47:0] <= partial_result_a[47:0] + partial_result_b[47:0];
end
//
// Now round unneed precision from accumulator result, and clip the unused dynamic range
//
wire [47:17] result_rnd;
round_reg #(.bits_in(48),.bits_out(31))
final_round (.clk(clk),.in(result[47:0]),.out(result_rnd[47:17]),.err());
wire [34:17] result_clip;
clip_reg #(.bits_in(31),.bits_out(18)) final_clip
(.clk(clk),.in(result_rnd[47:17]),.strobe_in(1'b1), .out(result_clip[34:17]), .strobe_out());
//
// Data enters the sample pipeline when stb_in is asserted.
// The clock cycle after is phase=0 with pipeline delay=0
//
always @(posedge clk)
if (rst) begin
stb_pipe0 <= 1'b0; // New sample loaded into sample pipeline, setup on reg a
stb_pipe1 <= 1'b0; // Sample n+0 presented to pre-adder, result to reg ad, coeff_a setup on input reg b
stb_pipe2 <= 1'b0; // Sample n+1 presented to pre-adder, product of sample n+0 and coeff_a setup on reg m
stb_pipe3 <= 1'b0; // product of sample n+1 and coeff_b setup on reg m, previous product loaded into accumulator.
stb_pipe4 <= 1'b0; // Add both products into accumulator
stb_pipe5 <= 1'b0; // Add 3 accumulator values into one partial result
stb_pipe6 <= 1'b0; // Add both partial results into final full precision result.
stb_pipe7 <= 1'b0; // Round result
stb_pipe8 <= 1'b0; // Clip Result
stb_pipe9 <= 1'b0;
end else begin
stb_pipe0 <= stb_in; // New sample loaded into sample pipeline, setup on reg a
stb_pipe1 <= stb_pipe0; // Sample n+0 presented to pre-adder, result to reg ad, coeff_a setup on input reg b
stb_pipe2 <= stb_pipe1; // Sample n+1 presented to pre-adder, product of sample n+0 and coeff_a setup on reg m
stb_pipe3 <= stb_pipe2; // product of sample n+1 and coeff_b setup on reg m, previous product loaded into accumulator.
stb_pipe4 <= stb_pipe3; // Add both products into accumulator
stb_pipe5 <= stb_pipe4; // Add 3 accumulator values into one partial result
stb_pipe6 <= stb_pipe5; // Add both partial results into final result.
stb_pipe7 <= stb_pipe6; // Round result
stb_pipe8 <= stb_pipe7; // Clip Result
stb_pipe9 <= stb_pipe8;
end // else: !if(rst)
//
// Interleave newly interpolated samples (odd taps) with raw input samples (even taps - All zero except center tap)
// Account for differences caused by various CPO settings and the pipeline advancing.
//
always @(posedge clk)
if (bypass)
data_out <= data_in;
else if (stb_in & stb_out)
data_out <= result_clip[34:17];
else if(stb_out)
case(output_rate)
1: data_out <= data_in_pipe[16]; // Four input pipeline shifts since we calculated odd taps
2: data_out <= data_in_pipe[14]; // Three input pipeline shifts since we calculated odd taps
3,4: data_out <= data_in_pipe[13]; // Two input pipeline shifts since we calculated odd taps
default: data_out <= data_in_pipe[12]; // One input pipeline shift since we calculated odd taps
endcase // case(output_rate)
endmodule // hb47_int
|
