Skip to content
This repository was archived by the owner on May 16, 2024. It is now read-only.

[ETConverter] Update text converter to new schema #78

Open
changhai0109 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[ETConverter] Update text converter to new schema #78

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 29 additions & 1 deletion et_converter/text2chakra_converter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
ALL_TO_ALL,
ALL_GATHER,
REDUCE_SCATTER,
GlobalMetadata
)

class Layer:
Expand Down Expand Up @@ -66,6 +67,17 @@ def __init__(
self.num_passes = num_passes
self.logger = logger
self.next_node_id = 0

def get_global_metadata(self):
input_text = ""
with open(self.input_filename, "r") as input_file:
input_text = input_file.read()
attr = [
ChakraAttr(name="schema", string_val="text2chakra_converter"),
ChakraAttr(name="input_file", string_val=input_text)
]
metadata = GlobalMetadata(attr=attr)
return metadata

def get_layers(
self,
Expand Down Expand Up @@ -126,14 +138,18 @@ def get_comm_coll_node(
node.attr.append(
ChakraAttr(name="comm_type",
int64_val=self.get_comm_type(comm_type)))
node.attr.append(
ChakraAttr(name="comm_size",
uint64_val = comm_size)
)
return node

def add_parent(
self,
child_node: Any,
parent_node: Any
) -> None:
child_node.parent.append(parent_node.id)
child_node.data_deps.append(parent_node.id)

def convert(self) -> None:
with open(self.input_filename, "r") as f:
Expand Down Expand Up @@ -167,6 +183,8 @@ def convert_microbenchmark(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
for layer in layers:
bwd_wg_comm_node = self.get_comm_coll_node(
Expand All @@ -190,6 +208,8 @@ def convert_data_parallel(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
fwd_comp_node = None

Expand Down Expand Up @@ -252,6 +272,8 @@ def convert_model_parallel(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
fwd_comm_node = None

Expand Down Expand Up @@ -327,6 +349,8 @@ def convert_hybrid_data_model(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
fwd_comm_node = None

Expand Down Expand Up @@ -416,6 +440,8 @@ def convert_hybrid_model_data(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
fwd_comm_node = None

Expand Down Expand Up @@ -504,6 +530,8 @@ def convert_hybrid_dlrm(
for npu_id in range(self.num_npus):
output_filename = "%s.%d.et" % (self.output_filename, npu_id)
with open(output_filename, "wb") as g:
global_metadata = self.get_global_metadata()
encode_message(g, global_metadata)
for i in range(self.num_passes):
fwd_comp_node = None

Expand Down