169 lines
6.3 KiB
Python
169 lines
6.3 KiB
Python
|
import torch
|
||
|
|
from cpp_to_py import io_parser
|
||
|
|
|
||
|
|
|
||
|
|
class IOParser(object):
|
||
|
|
def __init__(self) -> None:
|
||
|
|
super().__init__()
|
||
|
|
self.params = None
|
||
|
|
|
||
|
|
def load_params(self, params, verbose_log, lite_mode, random_place, num_threads):
|
||
|
|
check_status = self.check_params(params, verbose_log, lite_mode, random_place, num_threads)
|
||
|
|
if not check_status:
|
||
|
|
raise ValueError(
|
||
|
|
"Checking failure. Please check the validity of params: %s" % params
|
||
|
|
)
|
||
|
|
load_status = io_parser.load_params(self.params)
|
||
|
|
if not load_status:
|
||
|
|
raise ValueError(
|
||
|
|
"Loading failure. Please check the validity of params: %s" % params
|
||
|
|
)
|
||
|
|
return load_status
|
||
|
|
|
||
|
|
def check_params(
|
||
|
|
self, params: dict, verbose_log: bool, lite_mode: bool, random_place: bool, num_threads: int = 8
|
||
|
|
) -> bool:
|
||
|
|
if "def" not in params.keys() and "aux" not in params.keys():
|
||
|
|
print("def or aux is not found!")
|
||
|
|
return False
|
||
|
|
if "def" in params.keys() and "aux" in params.keys():
|
||
|
|
print("Can only support one input format!")
|
||
|
|
return False
|
||
|
|
if "def" in params.keys():
|
||
|
|
if "lef" not in params.keys():
|
||
|
|
if "cell_lef" not in params.keys() and "tech_lef" not in params.keys():
|
||
|
|
print("lef or (cell_lef and tech_lef) is not found")
|
||
|
|
return False
|
||
|
|
if "output" in params.keys():
|
||
|
|
if "def" != params["output"].split(".")[-1]:
|
||
|
|
print("output format should be .def")
|
||
|
|
return False
|
||
|
|
if "aux" in params.keys():
|
||
|
|
if "pl" not in params.keys():
|
||
|
|
print("pl is not found!")
|
||
|
|
return False
|
||
|
|
if "output" in params.keys():
|
||
|
|
if "pl" != params["output"].split(".")[-1]:
|
||
|
|
print("output format should be .pl")
|
||
|
|
return False
|
||
|
|
self.params = params
|
||
|
|
|
||
|
|
if verbose_log:
|
||
|
|
self.params["verbose_parser_log"] = True
|
||
|
|
else:
|
||
|
|
self.params["verbose_parser_log"] = False
|
||
|
|
|
||
|
|
if lite_mode:
|
||
|
|
self.params["lite_mode"] = True
|
||
|
|
else:
|
||
|
|
self.params["lite_mode"] = False
|
||
|
|
|
||
|
|
if random_place:
|
||
|
|
self.params["random_place"] = True
|
||
|
|
else:
|
||
|
|
self.params["random_place"] = False
|
||
|
|
|
||
|
|
self.params["num_threads"] = num_threads
|
||
|
|
|
||
|
|
return True
|
||
|
|
|
||
|
|
def read(
|
||
|
|
self,
|
||
|
|
params: dict,
|
||
|
|
verbose_log: bool = False,
|
||
|
|
lite_mode: bool = False,
|
||
|
|
random_place: bool = True,
|
||
|
|
num_threads: int = 8,
|
||
|
|
debug: bool = False,
|
||
|
|
):
|
||
|
|
check_status = self.check_params(params, verbose_log, lite_mode, random_place, num_threads)
|
||
|
|
if not check_status:
|
||
|
|
raise ValueError(
|
||
|
|
"Checking failure. Please check the validity of params: %s" % params
|
||
|
|
)
|
||
|
|
if debug:
|
||
|
|
load_status = io_parser.load_params(self.params)
|
||
|
|
if not load_status:
|
||
|
|
raise ValueError(
|
||
|
|
"Loading failure. Please check the validity of params: %s" % params
|
||
|
|
)
|
||
|
|
rawdb = io_parser.create_database()
|
||
|
|
rawdb.load()
|
||
|
|
rawdb.setup()
|
||
|
|
gpdb = io_parser.create_gpdatabase(rawdb)
|
||
|
|
gpdb.setup()
|
||
|
|
else:
|
||
|
|
rawdb, gpdb = io_parser.start(self.params)
|
||
|
|
# rawdb and gpdb are both c++ shared pointer
|
||
|
|
return (rawdb, gpdb)
|
||
|
|
|
||
|
|
def preprocess_design_info(self, gpdb):
|
||
|
|
dieLX, dieHX, dieLY, dieHY = gpdb.coreInfo() # use coreInfo instead of dieInfo
|
||
|
|
|
||
|
|
die_info = torch.tensor([dieLX, dieHX, dieLY, dieHY]).float()
|
||
|
|
# die_shift = torch.tensor([dieLX, dieLY])
|
||
|
|
# die_scale = torch.tensor([dieHX - dieLX, dieHY - dieLY])
|
||
|
|
|
||
|
|
siteWidth = gpdb.siteWidth()
|
||
|
|
siteHeight = gpdb.siteHeight()
|
||
|
|
site_info = (float(siteWidth), float(siteHeight))
|
||
|
|
|
||
|
|
node_pos = gpdb.node_cpos_tensor()
|
||
|
|
node_size = gpdb.node_size_tensor()
|
||
|
|
pin_rel_cpos = gpdb.pin_rel_cpos_tensor()
|
||
|
|
pin_size = gpdb.pin_size_tensor()
|
||
|
|
|
||
|
|
pin_id2node_id = gpdb.pin_id2node_id_tensor()
|
||
|
|
(
|
||
|
|
hyperedge_index,
|
||
|
|
hyperedge_list,
|
||
|
|
hyperedge_list_end,
|
||
|
|
) = gpdb.hyperedge_info_tensor()
|
||
|
|
|
||
|
|
node_id2region_id, region_boxes, region_boxes_end = gpdb.region_info_tensor()
|
||
|
|
|
||
|
|
node_type_indices = gpdb.node_type_indices()
|
||
|
|
node_id2node_name = gpdb.node_id2node_name()
|
||
|
|
all_node_types = []
|
||
|
|
mov_end_idx = None
|
||
|
|
fix_end_idx = None
|
||
|
|
connected_end_idx = None
|
||
|
|
for start_idx, end_idx, type_name in node_type_indices:
|
||
|
|
all_node_types.append(type_name)
|
||
|
|
if "FloatMov" == type_name:
|
||
|
|
mov_end_idx = end_idx
|
||
|
|
if "FloatFix" == type_name:
|
||
|
|
fix_end_idx = end_idx
|
||
|
|
if "IOPin" == type_name:
|
||
|
|
connected_end_idx = end_idx
|
||
|
|
# Mov + FloatMov
|
||
|
|
movable_index = (0, mov_end_idx)
|
||
|
|
# Mov + FloatMov + Fix + IOPin
|
||
|
|
connected_index = (0, connected_end_idx)
|
||
|
|
# Fix + IOPin + Blkg + FloatIOPin + FloatFix
|
||
|
|
fixed_index = (mov_end_idx, fix_end_idx)
|
||
|
|
|
||
|
|
design_info = {
|
||
|
|
"benchmark": self.params["benchmark"],
|
||
|
|
"dataset_path": self.params,
|
||
|
|
"node_type_indices": node_type_indices,
|
||
|
|
"node_id2node_name": node_id2node_name,
|
||
|
|
"movable_index": movable_index,
|
||
|
|
"connected_index": connected_index,
|
||
|
|
"fixed_index": fixed_index,
|
||
|
|
"site_info": site_info,
|
||
|
|
"die_info": die_info,
|
||
|
|
"node_pos": node_pos.contiguous(),
|
||
|
|
"node_size": node_size.contiguous(),
|
||
|
|
"pin_rel_cpos": pin_rel_cpos.contiguous(),
|
||
|
|
"pin_size": pin_size.contiguous(),
|
||
|
|
"pin_id2node_id": pin_id2node_id.long().contiguous(),
|
||
|
|
"hyperedge_index": hyperedge_index.long().contiguous(),
|
||
|
|
"hyperedge_list": hyperedge_list.long().contiguous(),
|
||
|
|
"hyperedge_list_end": hyperedge_list_end.long().contiguous(),
|
||
|
|
"node_id2region_id": node_id2region_id.long().contiguous(),
|
||
|
|
"region_boxes": region_boxes.contiguous(),
|
||
|
|
"region_boxes_end": region_boxes_end.long().contiguous(),
|
||
|
|
}
|
||
|
|
return design_info
|