-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathpe_k_gt30.vhd
547 lines (462 loc) · 17.6 KB
/
pe_k_gt30.vhd
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
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
-- File: pe_k_gt30.vhd
-- Date: 19 November 2009
-- Project: Zuker
-- Author: Arpith Chacko Jacob
-- Department of Computer Science and Engineering
-- Washington University in Saint Louis
-- Description:
-- Processing element implementing the cell k>30 of the zuker linear
-- array.
library ieee;
use ieee.std_logic_1164.all;
use ieee.std_logic_signed.all;
library work;
use work.zuker_pkg.all;
entity pe_k_gt30 is
generic (
-- When true delay registers are placed in block ram memories
USE_BRAM : boolean := false
);
port (
clk : in sl;
reset : in sl;
--
-- INPUT
-- DEPENDENCIES
--
-- T(i, j, k+1)
in_T_i_j_kp1 : in slv (CELL_WIDTH-1 downto 0);
-- PW2(i+1, j, k-1)
in_PW2_ip1_j_km1 : in slv (CELL_WIDTH-1 downto 0);
-- PW3(i, j-1, k-1)
in_PW3_i_jm1_km1 : in slv (CELL_WIDTH-1 downto 0);
-- CONTROL SIGNALS
--
in_k_eq_j_m_i_div_2 : in sl;
in_k2_eq_j_m_i : in sl;
--
-- OUTPUT
--
-- output dependencies
out_PW1_i_j_k : out slv (CELL_WIDTH-1 downto 0);
out_PW2_i_j_k : out slv (CELL_WIDTH-1 downto 0);
out_PW3_i_j_k : out slv (CELL_WIDTH-1 downto 0);
out_PW4_i_j_k : out slv (CELL_WIDTH-1 downto 0);
out_T_i_j_k : out slv (CELL_WIDTH-1 downto 0);
-- output control signals
out_k_eq_j_m_i_div_2 : out sl;
out_k2_eq_j_m_i : out sl
);
end entity;
architecture rtl of pe_k_gt30 is
component delay_pipe is
generic (
DATA_WIDTH : natural := 1;
DELAY_STATES : natural := 1;
NEEDS_RESET : boolean := true;
-- When needs reset=true, True: set all vals to 1, else 0
RESET_VAL_HIGH : boolean := false
);
port (
clk : in sl;
reset : in sl;
din : in slv (DATA_WIDTH-1 downto 0);
dout : out slv (DATA_WIDTH-1 downto 0)
);
end component;
component delay_fifo is
port (
rd_en : in STD_LOGIC := 'X';
wr_en : in STD_LOGIC := 'X';
full : out STD_LOGIC;
empty : out STD_LOGIC;
clk : in STD_LOGIC := 'X';
srst : in STD_LOGIC := 'X';
dout : out STD_LOGIC_VECTOR ( 17 downto 0 );
din : in STD_LOGIC_VECTOR ( 17 downto 0 )
);
end component;
component pw1
port (
PW1_in : in slv (CELL_WIDTH-1 downto 0);
PW3 : in slv (CELL_WIDTH-1 downto 0);
k2_eq_j_m_i : in sl;
PW1 : out slv (CELL_WIDTH-1 downto 0)
);
end component;
component pw2_kgt1 is
port (
PW2_in : in slv (CELL_WIDTH-1 downto 0);
PW2 : out slv (CELL_WIDTH-1 downto 0)
);
end component;
component pw3_kgt1 is
port (
PW3_in : in slv (CELL_WIDTH-1 downto 0);
PW3 : out slv (CELL_WIDTH-1 downto 0)
);
end component;
component pw4 is
port (
PW4_in : in slv (CELL_WIDTH-1 downto 0);
PW2 : in slv (CELL_WIDTH-1 downto 0);
k2_eq_j_m_i : in sl;
PW4 : out slv (CELL_WIDTH-1 downto 0)
);
end component;
component t is
port (
T_in : in slv (CELL_WIDTH-1 downto 0);
PW1 : in slv (CELL_WIDTH-1 downto 0);
PW2 : in slv (CELL_WIDTH-1 downto 0);
PW3 : in slv (CELL_WIDTH-1 downto 0);
PW4 : in slv (CELL_WIDTH-1 downto 0);
k_eq_j_m_i_div_2 : in sl;
T : out slv (CELL_WIDTH-1 downto 0)
);
end component;
-----------------------------------------------------------------------------
-- Link signals
-----------------------------------------------------------------------------
signal int_k2_eq_j_m_i : slv (0 downto 0);
signal int_k_eq_j_m_i_div_2 : slv (0 downto 0);
signal in_k2_eq_j_m_i_link : slv (0 downto 0);
signal in_k_eq_j_m_i_div_2_link : slv (0 downto 0);
signal in_T_i_j_kp1_link : slv (CELL_WIDTH-1 downto 0);
signal in_PW1_i_jm1_k_link : slv (CELL_WIDTH-1 downto 0);
signal in_PW2_ip1_j_km1_link : slv (CELL_WIDTH-1 downto 0);
signal in_PW3_i_jm1_km1_link : slv (CELL_WIDTH-1 downto 0);
signal in_PW4_ip1_j_k_link : slv (CELL_WIDTH-1 downto 0);
-----------------------------------------------------------------------------
-- Signals used for block ram based delays
-----------------------------------------------------------------------------
-- signals used for FIFO delay
signal wr_start : slv(0 downto 0);
signal delay_wr : slv(0 downto 0);
signal delay_fifo_din1 : slv(17 downto 0);
signal delay_fifo_rd1 : slv(0 downto 0);
signal delay_fifo_dout1 : slv(17 downto 0);
signal delay_fifo_din2 : slv(17 downto 0);
signal delay_fifo_rd2 : slv(0 downto 0);
signal delay_fifo_dout2 : slv(17 downto 0);
type RETIMING_ARRAY is array (1 to NO_RETIMING_REGS) of slv(17 downto 0);
signal fifo_retiming1 : RETIMING_ARRAY;
signal fifo_retiming2 : RETIMING_ARRAY;
-----------------------------------------------------------------------------
-- Intermediate computation signals
-----------------------------------------------------------------------------
signal PW1_i_j_k : slv (CELL_WIDTH-1 downto 0);
signal PW2_i_j_k : slv (CELL_WIDTH-1 downto 0);
signal PW3_i_j_k : slv (CELL_WIDTH-1 downto 0);
signal PW4_i_j_k : slv (CELL_WIDTH-1 downto 0);
signal T_i_j_k : slv (CELL_WIDTH-1 downto 0);
-----------------------------------------------------------------------------
-- Output signals
-----------------------------------------------------------------------------
signal PW1_i_j_k_reg : slv (CELL_WIDTH-1 downto 0);
signal PW2_i_j_k_reg : slv (CELL_WIDTH-1 downto 0);
signal PW3_i_j_k_reg : slv (CELL_WIDTH-1 downto 0);
signal PW4_i_j_k_reg : slv (CELL_WIDTH-1 downto 0);
signal T_i_j_k_reg : slv (CELL_WIDTH-1 downto 0);
begin
-----------------------------------------------------------------------------
-- Delay control pipeline communication links according to the schedule
-----------------------------------------------------------------------------
-- in_k2_eq_j_m_i(i+2, j, k-1) -- to initialize pipelines PW1, PW2
int_k2_eq_j_m_i(0) <= in_k2_eq_j_m_i;
delay_k2_eq_j_m_i : delay_pipe
generic map (
DATA_WIDTH => 1,
DELAY_STATES => 2, -- delay = 3
NEEDS_RESET => NEEDS_RESET,
RESET_VAL_HIGH => false
)
port map (
clk => clk,
reset => reset,
din => int_k2_eq_j_m_i,
dout => in_k2_eq_j_m_i_link
);
-- in_k_eq_j_m_i_div_2(i+2, j, k-1) -- to start aggregation at k = (j-i)/2
int_k_eq_j_m_i_div_2(0) <= in_k_eq_j_m_i_div_2;
delay_k_eq_j_m_i_div_2 : delay_pipe
generic map (
DATA_WIDTH => 1,
DELAY_STATES => 2, -- delay = 3
NEEDS_RESET => NEEDS_RESET,
RESET_VAL_HIGH => false
)
port map (
clk => clk,
reset => reset,
din => int_k_eq_j_m_i_div_2,
dout => in_k_eq_j_m_i_div_2_link
);
-----------------------------------------------------------------------------
-- Delay communication links according to the schedule
-----------------------------------------------------------------------------
-- T(i, j, k+1): delay by 1 clock cycle. 1 clock cycle when output is
-- registered at source, so no need to register anymore
in_T_i_j_kp1_link <= in_T_i_j_kp1;
-- PW2(i+1, j, k-1): delay by 1 clock cycle. 1 clock cycle when output is
-- registered at source, so no need to register anymore
in_PW2_ip1_j_km1_link <= in_PW2_ip1_j_km1;
-- PW4(i+1, j, k): delay by 2 clock cycles
delay_PW4_ip1_j_k : delay_pipe
generic map (
DATA_WIDTH => CELL_WIDTH,
DELAY_STATES => 1, -- delay = 2
NEEDS_RESET => NEEDS_RESET,
RESET_VAL_HIGH => false
)
port map (
clk => clk,
reset => reset,
din => PW4_i_j_k_reg,
dout => in_PW4_ip1_j_k_link
);
-----------------------------------------------------------------------------
-- Generate memory intensive link delays using LUTS
-----------------------------------------------------------------------------
gen_delays_luts : if (not USE_BRAM) generate
-- PW1(i, j-1, k): delay by N clock cycles
delay_PW1_i_jm1_k : delay_pipe
generic map (
DATA_WIDTH => CELL_WIDTH,
DELAY_STATES => SEQ_LEN-1, -- delay = SEQ_LEN
NEEDS_RESET => NEEDS_RESET,
RESET_VAL_HIGH => false
)
port map (
clk => clk,
reset => reset,
din => PW1_i_j_k_reg,
dout => in_PW1_i_jm1_k_link
);
-- PW3(i, j-1, k-1): delay by N-1 clock cycles
delay_PW3_i_jm1_km1 : delay_pipe
generic map (
DATA_WIDTH => CELL_WIDTH,
DELAY_STATES => SEQ_LEN-2, -- delay = SEQ_LEN-1
NEEDS_RESET => NEEDS_RESET,
RESET_VAL_HIGH => false
)
port map (
clk => clk,
reset => reset,
din => in_PW3_i_jm1_km1,
dout => in_PW3_i_jm1_km1_link
);
end generate gen_delays_luts;
-----------------------------------------------------------------------------
-- Generate memory intensive link delays using block RAMs
-----------------------------------------------------------------------------
gen_delays_bram : if (USE_BRAM) generate
-- delay of SEQ_LEN should not exceed capacity of block ram which is
-- 1024.
assert (SEQ_LEN <= 1024/2) report
"pe_k_2 block ram has insufficient capacity" severity error;
---------------------------------------------------------------------------
-- PW1(i, j-1, k): delay by N clock cycles
---------------------------------------------------------------------------
-- input into delay block ram
delay_fifo_din1 <= "00" & PW1_i_j_k_reg;
-- instantiates block ram
delay_link1 : delay_fifo
port map (
clk => clk,
srst => reset,
rd_en => delay_fifo_rd1(0),
dout => delay_fifo_dout1,
full => open,
wr_en => delay_wr(0),
din => delay_fifo_din1,
empty => open
);
-- delay element to read from fifo at correct time
fifo_rd1 : delay_pipe
generic map (
DATA_WIDTH => 1,
DELAY_STATES => SEQ_LEN-2-NO_RETIMING_REGS, -- delay = SEQ_LEN
NEEDS_RESET => true, -- 1 clock at output of PE
RESET_VAL_HIGH => false -- 1 clock latency to read the fifo and
) -- NO_RETIMING_REGS clocks to improve timing
port map (
clk => clk,
reset => reset,
din => delay_wr,
dout => delay_fifo_rd1
);
-- output from delay block ram is registered multiple clocks to help
-- improve timing using retiming.
reg_output_delay1 : process (clk)
begin
if rising_edge(clk) then
fifo_retiming1(1) <= delay_fifo_dout1;
for i in 2 to NO_RETIMING_REGS loop
fifo_retiming1(i) <= fifo_retiming1(i-1);
end loop;
end if;
end process reg_output_delay1;
-- output: data from this link goes into PE logic
-- PW1(i, j-1, k): delay by N clock cycles
in_PW1_i_jm1_k_link <= fifo_retiming1(NO_RETIMING_REGS)
(CELL_WIDTH-1 downto 0);
---------------------------------------------------------------------------
-- PW3(i, j-1, k-1): delay by N-1 clock cycles
---------------------------------------------------------------------------
-- input into delay block ram
delay_fifo_din2 <= "00" & in_PW3_i_jm1_km1;
-- instantiates block ram
delay_link2 : delay_fifo
port map (
clk => clk,
srst => reset,
rd_en => delay_fifo_rd2(0),
dout => delay_fifo_dout2,
full => open,
wr_en => delay_wr(0),
din => delay_fifo_din2,
empty => open
);
-- delay element to read from fifo at correct time
fifo_rd2 : delay_pipe
generic map (
DATA_WIDTH => 1,
DELAY_STATES => SEQ_LEN-3-NO_RETIMING_REGS, -- delay = SEQ_LEN-1
NEEDS_RESET => true, -- 1 clock at output of PE
RESET_VAL_HIGH => false -- 1 clock latency to read the fifo and
) -- NO_RETIMING_REGS clocks to improve timing
port map (
clk => clk,
reset => reset,
din => delay_wr,
dout => delay_fifo_rd2
);
-- output from delay block ram is registered multiple clocks to help
-- improve timing using retiming.
reg_output_delay2 : process (clk)
begin
if rising_edge(clk) then
fifo_retiming2(1) <= delay_fifo_dout2;
for i in 2 to NO_RETIMING_REGS loop
fifo_retiming2(i) <= fifo_retiming2(i-1);
end loop;
end if;
end process reg_output_delay2;
-- output: data from this link goes into PE logic
-- PW3(i, j-1, k-1): delay by N-1 clock cycles
in_PW3_i_jm1_km1_link <= fifo_retiming2(NO_RETIMING_REGS)
(CELL_WIDTH-1 downto 0);
---------------------------------------------------------------------------
-- delay element to write to fifos a few clocks after reset is asserted
---------------------------------------------------------------------------
delay_fifo_wr : delay_pipe
generic map (
DATA_WIDTH => 1,
DELAY_STATES => 5, -- delay = 5 clocks
NEEDS_RESET => true, -- wait for a few clocks so that block
RESET_VAL_HIGH => false -- ram has a chance to settle
)
port map (
clk => clk,
reset => reset,
din => wr_start,
dout => delay_wr
);
-- registers to control reading and writing into block ram
reg_write_fifos : process (clk)
begin
if rising_edge(clk) then
if reset = '1' then
wr_start <= (others => '0');
else
-- keep writing into block ram based delay register
wr_start(0) <= '1';
end if;
end if; -- end rising_edge
end process reg_write_fifos;
end generate gen_delays_bram;
-----------------------------------------------------------------------------
-- Data variable PW1
-----------------------------------------------------------------------------
variable_pw1 : pw1
port map (
PW1_in => in_PW1_i_jm1_k_link,
PW3 => PW3_i_j_k,
k2_eq_j_m_i => in_k2_eq_j_m_i_link(0),
PW1 => PW1_i_j_k
);
-----------------------------------------------------------------------------
-- Data variable PW2
-----------------------------------------------------------------------------
variable_pw2 : pw2_kgt1
port map (
PW2_in => in_PW2_ip1_j_km1_link,
PW2 => PW2_i_j_k
);
-----------------------------------------------------------------------------
-- Data variable PW3
-----------------------------------------------------------------------------
variable_pw3 : pw3_kgt1
port map (
PW3_in => in_PW3_i_jm1_km1_link,
PW3 => PW3_i_j_k
);
-----------------------------------------------------------------------------
-- Data variable PW4
-----------------------------------------------------------------------------
variable_pw4 : pw4
port map (
PW4_in => in_PW4_ip1_j_k_link,
PW2 => PW2_i_j_k,
k2_eq_j_m_i => in_k2_eq_j_m_i_link(0),
PW4 => PW4_i_j_k
);
-----------------------------------------------------------------------------
-- Data variable T
-----------------------------------------------------------------------------
variable_t : t
port map (
T_in => in_T_i_j_kp1_link,
PW1 => PW1_i_j_k,
PW2 => PW2_i_j_k,
PW3 => PW3_i_j_k,
PW4 => PW4_i_j_k,
k_eq_j_m_i_div_2 => in_k_eq_j_m_i_div_2_link(0),
T => T_i_j_k
);
-----------------------------------------------------------------------------
-- Process to register computed values for data variables
-----------------------------------------------------------------------------
reg_values_noreset: process (clk)
begin
if rising_edge(clk) then
-- register computed values: output
PW1_i_j_k_reg <= PW1_i_j_k;
PW2_i_j_k_reg <= PW2_i_j_k;
PW3_i_j_k_reg <= PW3_i_j_k;
PW4_i_j_k_reg <= PW4_i_j_k;
T_i_j_k_reg <= T_i_j_k;
end if; -- end rising_edge
end process reg_values_noreset;
reg_values: process (clk)
begin
if rising_edge(clk) then
if reset = '1' then
-- register control signals
out_k_eq_j_m_i_div_2 <= '0';
out_k2_eq_j_m_i <= '0';
else
-- register control signals
out_k_eq_j_m_i_div_2 <= in_k_eq_j_m_i_div_2_link(0);
out_k2_eq_j_m_i <= in_k2_eq_j_m_i_link(0);
end if;
end if; -- end rising_edge
end process reg_values;
out_PW1_i_j_k <= PW1_i_j_k_reg;
out_PW2_i_j_k <= PW2_i_j_k_reg;
out_PW3_i_j_k <= PW3_i_j_k_reg;
out_PW4_i_j_k <= PW4_i_j_k_reg;
out_T_i_j_k <= T_i_j_k_reg;
end rtl;