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
|
# SPDX-FileCopyrightText: Copyright 2022-2024, Arm Limited and/or its affiliates.
# SPDX-License-Identifier: Apache-2.0
"""Module for backend integration."""
from __future__ import annotations
import base64
import json
import logging
import re
import subprocess # nosec
from dataclasses import dataclass
from pathlib import Path
from mlia.backend.errors import BackendExecutionFailed
from mlia.backend.repo import get_backend_repository
from mlia.utils.filesystem import get_mlia_resources
from mlia.utils.proc import Command
from mlia.utils.proc import OutputLogger
from mlia.utils.proc import process_command_output
logger = logging.getLogger(__name__)
@dataclass
class PerformanceMetrics:
"""Performance metrics parsed from generic inference output."""
npu_active_cycles: int
npu_idle_cycles: int
npu_total_cycles: int
npu_axi0_rd_data_beat_received: int
npu_axi0_wr_data_beat_written: int
npu_axi1_rd_data_beat_received: int
class GenericInferenceOutputParser:
"""Generic inference runner output parser."""
pattern = re.compile(r"<metrics>(.*)</metrics>")
def __init__(self) -> None:
"""Init parser."""
self.base64_data: list[str] = []
def __call__(self, line: str) -> None:
"""Extract base64 strings from the app output."""
if res_b64 := self.pattern.search(line):
self.base64_data.append(res_b64.group(1))
def get_metrics(self) -> PerformanceMetrics:
"""Parse the collected data and return perf metrics."""
try:
parsed_metrics = self._parse_data()
return PerformanceMetrics(
parsed_metrics["NPU ACTIVE"],
parsed_metrics["NPU IDLE"],
parsed_metrics["NPU TOTAL"],
parsed_metrics["NPU AXI0_RD_DATA_BEAT_RECEIVED"],
parsed_metrics["NPU AXI0_WR_DATA_BEAT_WRITTEN"],
parsed_metrics["NPU AXI1_RD_DATA_BEAT_RECEIVED"],
)
except Exception as err:
raise ValueError("Unable to parse output and get metrics.") from err
def _parse_data(self) -> dict[str, int]:
"""Parse the data."""
parsed_metrics: dict[str, int] = {}
for base64_item in self.base64_data:
res_json = base64.b64decode(base64_item, validate=True)
for profiling_group in json.loads(res_json):
for metric in profiling_group["samples"]:
metric_name = metric["name"]
metric_value = int(metric["value"][0])
if metric_name in parsed_metrics:
raise KeyError(f"Duplicate key {metric_name}")
parsed_metrics[metric_name] = metric_value
return parsed_metrics
@dataclass
class FVPMetadata:
"""Metadata for FVP."""
executable: str
generic_inf_app: Path
def get_generic_inference_app_path(fvp: str, target: str) -> Path:
"""Return path to the generic inference runner binary."""
apps_path = get_mlia_resources() / "backends/applications"
fvp_mapping = {"corstone-300": "300", "corstone-310": "310"}
target_mapping = {"ethos-u55": "U55", "ethos-u65": "U65"}
fvp_version = f"sse-{fvp_mapping[fvp]}"
app_version = f"22.08.02-ethos-{target_mapping[target]}-Default-noTA"
app_dir = f"inference_runner-{fvp_version}-{app_version}"
return apps_path.joinpath(app_dir, "ethos-u-inference_runner.axf")
def get_executable_name(fvp: str, profile: str, target: str) -> str:
"""Return name of the executable for selected FVP and profile."""
executable_name_mapping = {
("corstone-300", "AVH", "ethos-u55"): "VHT_Corstone_SSE-300_Ethos-U55",
("corstone-300", "AVH", "ethos-u65"): "VHT_Corstone_SSE-300_Ethos-U65",
("corstone-300", "default", "ethos-u55"): "FVP_Corstone_SSE-300_Ethos-U55",
("corstone-300", "default", "ethos-u65"): "FVP_Corstone_SSE-300_Ethos-U65",
("corstone-310", "AVH", "ethos-u55"): "VHT_Corstone_SSE-310",
("corstone-310", "AVH", "ethos-u65"): "VHT_Corstone_SSE-310_Ethos-U65",
("corstone-310", "default", "ethos-u55"): "FVP_Corstone_SSE-310",
("corstone-310", "default", "ethos-u65"): "FVP_Corstone_SSE-310_Ethos-U65",
}
return executable_name_mapping[(fvp, profile, target)]
def get_fvp_metadata(fvp: str, profile: str, target: str) -> FVPMetadata:
"""Return metadata for selected Corstone backend."""
executable_name = get_executable_name(fvp, profile, target)
app = get_generic_inference_app_path(fvp, target)
return FVPMetadata(executable_name, app)
def build_corstone_command(
backend_path: Path,
fvp: str,
target: str,
mac: int,
model: Path,
profile: str,
) -> Command:
"""Build command to run Corstone FVP."""
fvp_metadata = get_fvp_metadata(fvp, profile, target)
cmd = [
backend_path.joinpath(fvp_metadata.executable).as_posix(),
"-a",
fvp_metadata.generic_inf_app.as_posix(),
"--data",
f"{model}@0x90000000",
"-C",
f"ethosu.num_macs={mac}",
"-C",
"mps3_board.telnetterminal0.start_telnet=0",
"-C",
"mps3_board.uart0.out_file='-'",
"-C",
"mps3_board.uart0.shutdown_on_eot=1",
"-C",
"mps3_board.visualisation.disable-visualisation=1",
"--stat",
]
return Command(cmd)
def get_metrics(
backend_path: Path,
fvp: str,
target: str,
mac: int,
model: Path,
profile: str = "default",
) -> PerformanceMetrics:
"""Run generic inference and return perf metrics."""
try:
command = build_corstone_command(
backend_path,
fvp,
target,
mac,
model,
profile,
)
except Exception as err:
raise BackendExecutionFailed(
f"Unable to construct a command line for {fvp}"
) from err
output_parser = GenericInferenceOutputParser()
output_logger = OutputLogger(logger)
try:
process_command_output(
command,
[output_parser, output_logger],
)
except subprocess.CalledProcessError as err:
raise BackendExecutionFailed("Backend execution failed.") from err
return output_parser.get_metrics()
def estimate_performance(
target: str, mac: int, model: Path, backend: str
) -> PerformanceMetrics:
"""Get performance estimations."""
backend_repo = get_backend_repository()
backend_path, settings = backend_repo.get_backend_settings(backend)
if not settings or "profile" not in settings:
raise BackendExecutionFailed(f"Unable to configure backend {backend}.")
return get_metrics(
backend_path,
backend,
target,
mac,
model,
settings["profile"],
)
|
