AI4EDA-LaMPlace/04_lamplace.py at main · MIRALab-USTC/AI4EDA-LaMPlace · GitHub

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
import argparse
import os
import random
import time
import numpy as np
from common import grid_setting, my_inf
from place_db import PlaceDB
from utils import (
    Record,
    random_guiding,
    write_final_placement,
    write_pl_for_gp,
    draw_macros,
    get_l_flow,
    l_mask_placer,
    rank_macros,
    draw_macro_placement,
)

def set_seed(seed: int):
    np.random.seed(seed)
    random.seed(seed)


def swap(node1: Record, node2: Record, grid_size: int):
    node1.bottom_left_x, node2.bottom_left_x = node2.bottom_left_x, node1.bottom_left_x
    node1.bottom_left_y, node2.bottom_left_y = node2.bottom_left_y, node1.bottom_left_y

    node1.grid_x, node2.grid_x = node2.grid_x, node1.grid_x
    node1.grid_y, node2.grid_y = node2.grid_y, node1.grid_y

    node1.refresh(grid_size)
    node2.refresh(grid_size)


def place(
    init_round,
    stop_round,
    placedb: PlaceDB,
    grid_num,
    grid_size,
    placement_file,
    result_dir,
    fig_dir,
):

    place_records = []
    l_flow,lf = get_l_flow(placedb)
    m2m_flow = None
    node_id_ls = rank_macros(placedb,l_flow)

    best_value = my_inf
    for cnt in range(init_round):
        print(f"init {cnt}")
        place_record = random_guiding(node_id_ls, placedb, grid_size,grid_num)
        placed_macros, value = l_mask_placer(
            node_id_ls, placedb, grid_num, grid_size, place_record,l_flow,lf
        )
        if not value in [r[1] for r in place_records]:
            placement_file = os.path.join(result_dir, f"placement_{value:.3f}.csv")
            write_final_placement(placed_macros, value, placement_file)
            place_records.append((placed_macros,value))
        if value < best_value:
            best_value = value
            best_placed_macro = placed_macros
    print("begin ea")
    candidates = sorted(placedb.macro_name)
    if best_value != my_inf:
        place_record = best_placed_macro
        for cnt in range(stop_round):
            print(cnt)
            node_a = random.sample(rank_macros(placedb,l_flow)[:5],1)[0]
            node_b = random.sample(candidates, 1)[0]
            swap(place_record[node_a], place_record[node_b], grid_size)
            placed_macros, value = l_mask_placer(
            node_id_ls, placedb, grid_num, grid_size, place_record,l_flow,lf
        )
            # save frame for this EA iteration (attempt)
            try:
                frame_path = os.path.join(fig_dir, f"ea_{cnt:04d}_{value:.3f}.png")
                draw_macro_placement(placed_macros, frame_path, placedb, draw_id=False)
            except Exception as e:
                print(f"warn: failed to draw ea frame {cnt}: {e}")
            if value >= best_value:
                swap(place_record[node_a], place_record[node_b], grid_size)
            else:
                best_value= value
                best_placed_macro = place_record = placed_macros
            if not value in [r[1] for r in place_records]:
                placement_file = os.path.join(result_dir, f"placement_{value:.3f}.csv")
                write_final_placement(placed_macros, value, placement_file)
                place_records.append((placed_macros,value))
    return place_records


def main():
    parser = argparse.ArgumentParser(description="argparse testing")
    parser.add_argument("--benchmark", default= 'superblue1')
    parser.add_argument("--seed", default= 2024)
    parser.add_argument("--init_round", default=50)
    parser.add_argument("--stop_round", default=20)
    args = parser.parse_args()
    benchmark = args.benchmark
    seed1 = args.seed
    stop_round = int(args.stop_round)
    init_round = int(args.init_round)
    set_seed(int(seed1))
    grid_num = grid_setting[benchmark]["grid_num"]
    grid_size = grid_setting[benchmark]["grid_size"]
    placedb = PlaceDB(benchmark, grid_size)
    result_dir = os.path.join("placement_data/lamplace", f"{benchmark}")
    if not os.path.exists(result_dir):
        os.makedirs(result_dir)
    if not os.path.exists(os.path.join(result_dir, f"pl_{seed1}")):
        os.makedirs(os.path.join(result_dir, f"pl_{seed1}"))
    if not os.path.exists(os.path.join(result_dir, f"fig_{seed1}")):
        os.makedirs(os.path.join(result_dir, f"fig_{seed1}"))
    placement_file = os.path.join(result_dir, f"placement_seed_{seed1}.csv")
    fig_dir = os.path.join(result_dir, f"fig_{seed1}")
    start = time.time()
    place_records = place(
        init_round, stop_round, placedb, grid_num, grid_size, placement_file,result_dir,fig_dir
    )
    end = time.time()
    print(f"time: {end-start}s")
    place_records = sorted(place_records, key=lambda x: x[1])
    #record best 5 placements
    for i in range(5):
        place_record,value = place_records[i]
        pic_file = os.path.join(result_dir, f"fig_{seed1}/{benchmark}_{value:.3f}.png")
        placement_file = os.path.join(result_dir, f"placement_{value:.3f}.csv")
        draw_macros(placedb, placement_file, grid_size, pic_file)
        pl_file = os.path.join(result_dir, f"pl_{seed1}/{benchmark}_{value:.3f}.gp.pl")
        write_pl_for_gp(place_record, placedb, pl_file)


if __name__ == "__main__":
    main()