1
0
mirror of https://github.com/ARM-software/workload-automation.git synced 2025-03-14 06:38:36 +00:00

workloads/geekbench: Add support for Geekbench command-line build

Add Geekbench command-line build workload for Android targets.
Geekbench apks allow to user to run the tests altogether. Using the
command-line, a single test or multiple tests can be specified.

Signed-off-by: Elif Topuz <elif.topuz@arm.com>
This commit is contained in:
Elif Topuz 2025-03-03 10:13:23 +00:00 committed by Marc Bonnici
parent b03f28d1d5
commit 0732fa9cf0

View File

@ -1,4 +1,4 @@
# Copyright 2013-2018 ARM Limited
# Copyright 2013-2025 ARM Limited
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
@ -20,10 +20,11 @@ import tempfile
import json
from collections import defaultdict
from wa import ApkUiautoWorkload, Parameter
from wa import Workload, ApkUiautoWorkload, Parameter
from wa.framework.exception import ConfigError, WorkloadError
from wa.utils.misc import capitalize
from wa.utils.types import version_tuple
from wa.utils.types import version_tuple, list_or_integer
from wa.utils.exec_control import once
class Geekbench(ApkUiautoWorkload):
@ -370,3 +371,233 @@ class GeekbenchCorproate(Geekbench): # pylint: disable=too-many-ancestors
def namemify(basename, i):
return basename + (' {}'.format(i) if i else '')
class GeekbenchCmdline(Workload):
name = "geekbench_cli"
description = "Workload for running command line version Geekbench"
gb6_workloads = {
# Single-Core and Multi-Core
101: 'File Compression',
102: 'Navigation',
103: 'HTML5 Browser',
104: 'PDF Renderer',
105: 'Photo Library',
201: 'Clang',
202: 'Text Processing',
203: 'Asset Compression',
301: 'Object Detection',
402: 'Object Remover',
403: 'HDR',
404: 'Photo Filter',
501: 'Ray Tracer',
502: 'Structure from Motion',
# OpenCL and Vulkan
303: 'Face Detection',
406: 'Edge Detection',
407: 'Gaussian Blur',
503: 'Feature Matching',
504: 'Stereo Matching',
601: 'Particle Physics',
# Single-Core, Multi-Core, OpenCL, and Vulkan
302: 'Background Blur',
401: 'Horizon Detection',
}
gb5_workloads = {
# Single-Core and Multi-Core
101: 'AES-XTS',
201: 'Text Compression',
202: 'Image Compression',
203: 'Navigation',
204: 'HTML5',
205: 'SQLite',
206: 'PDF Rendering',
207: 'Text Rendering',
208: 'Clang',
209: 'Camera',
301: 'N-Body Physics',
302: 'Rigid Body Physics',
307: 'Image Inpainting',
308: 'HDR',
309: 'Ray Tracing',
310: 'Structure from Motion',
312: 'Speech Recognition',
313: 'Machine Learning',
# OpenCL and Vulkan
220: 'Sobel',
221: 'Canny',
222: 'Stereo Matching',
230: 'Histogram Equalization',
304: 'Depth of Field',
311: 'Feature Matching',
320: 'Particle Physics',
321: 'SFFT',
# Single-Core, Multi-Core, OpenCL, and Vulkan
303: 'Gaussian Blur',
305: 'Face Detection',
306: 'Horizon Detection',
}
binary_name = 'geekbench_aarch64'
allowed_extensions = ['json', 'csv', 'xml', 'html', 'text']
parameters = [
Parameter('cpumask', kind=str, default='',
description='CPU mask used by taskset.'),
Parameter('section', kind=int, default=1, allowed_values=[1, 4, 9],
description="""Run the specified sections. It should be 1 for CPU benchmarks,
4 for OpenCL benchmarks and 9 for Vulkan benchmarks."""),
Parameter('upload', kind=bool, default=False,
description='Upload results to Geekbench Browser'),
Parameter('is_single_core', kind=bool, default=True,
description='Run workload in single-core or multi-core mode.'),
Parameter('workload', kind=list_or_integer, default=301,
description='Specify workload to run'),
Parameter('iterations', kind=int, default=5,
description='Number of iterations'),
Parameter('workload_gap', kind=int, default=2000,
description='N milliseconds gap between workloads'),
Parameter('output_file', kind=str, default='gb_cli.json',
description=f"""Specify the name of the output results file.
If it is not specified, the output file will be generated as a JSON file.
It can be {', '.join(allowed_extensions)} files."""),
Parameter('timeout', kind=int, default=2000,
description='The test timeout in ms. It should be long for 1000 iterations.'),
Parameter('version', kind=str, default='6.3.0',
description='Specifies which version of the Geekbench should run.'),
]
def __init__(self, target, **kwargs):
super(GeekbenchCmdline, self).__init__(target, **kwargs)
self.target_result_json = None
self.host_result_json = None
self.workloads = self.gb6_workloads
self.params = ''
self.output = ''
self.target_exec_directory = ''
self.tar_file_src = ''
self.tar_file_dst = ''
self.file_exists = False
def init_resources(self, context):
"""
Retrieves necessary files to run the benchmark in TAR format.
WA will look for `gb_cli_artifacts_<version>.tar` file to deploy them to the
working directory. If there is no specified version, it will look for version
6.3.0 by default.
"""
self.deployable_assets = [''.join(['gb_cli_artifacts', '_', self.version, '.tar'])]
# Create an executables directory
self.target_exec_directory = self.target.path.join(self.target.executables_directory, f'gb_cli-{self.version}')
self.target.execute("mkdir -p {}".format(self.target_exec_directory))
# Source and Destination paths for the artifacts tar file
self.tar_file_src = self.target.path.join(self.target.working_directory, self.deployable_assets[0])
self.tar_file_dst = self.target.path.join(self.target_exec_directory, self.deployable_assets[0])
# Check the tar file if it already exists
if self.target.file_exists(self.tar_file_dst):
self.file_exists = True
else:
# Get the assets file
super(GeekbenchCmdline, self).init_resources(context)
@once
def initialize(self, context):
if self.version[0] == '5':
self.workloads = self.gb5_workloads
# If the tar file does not exist in the target, deploy the assets
if not self.file_exists:
super(GeekbenchCmdline, self).initialize(context)
# Move the tar file to the executables directory
self.target.execute(
'{} mv {} {}'.format(
self.target.busybox, self.tar_file_src, self.tar_file_dst))
# Extract the tar file
self.target.execute(
'{} tar -xf {} -C {}'.format(
self.target.busybox, self.tar_file_dst, self.target_exec_directory))
def setup(self, context):
super(GeekbenchCmdline, self).setup(context)
self.params = ''
self.params += '--section {} '.format(self.section)
if self.section == 1:
self.params += '--single-core ' if self.is_single_core else '--multi-core '
self.params += '--upload ' if self.upload else '--no-upload '
known_workloads = '\n'.join("{}: {}".format(k, v) for k, v in self.workloads.items())
if any([t not in self.workloads.keys() for t in self.workload]):
msg = 'Unknown workload(s) specified. Known workloads: {}'
raise ValueError(msg.format(known_workloads))
self.params += '--workload {} '.format(''.join("{},".format(i) for i in self.workload))
if self.iterations:
self.params += '--iterations {} '.format(self.iterations)
if self.workload_gap:
self.params += '--workload-gap {} '.format(self.workload_gap)
extension = os.path.splitext(self.output_file)[1][1:]
if self.output_file and extension not in self.allowed_extensions:
msg = f"No allowed extension specified. Allowed extensions: {', '.join(self.allowed_extensions)}"
raise ValueError(msg)
elif self.output_file:
# Output results file with the given name and extension
self.target_result_json = os.path.join(self.target_exec_directory, self.output_file)
self.params += '--export-{} {}'.format(extension, self.target_result_json)
self.host_result_json = os.path.join(context.output_directory, self.output_file)
else:
# The output file is not specified
self.target_result_json = os.path.join(self.target_exec_directory, self.output_file)
self.params += '--save {}'.format(self.target_result_json)
self.host_result_json = os.path.join(context.output_directory, self.output_file)
def run(self, context):
super(GeekbenchCmdline, self).run(context)
taskset = f"taskset {self.cpumask}" if self.cpumask else ""
binary = self.target.path.join(self.target_exec_directory, self.binary_name)
cmd = '{} {} {}'.format(taskset, binary, self.params)
try:
self.output = self.target.execute(cmd, timeout=self.timeout, as_root=True)
except KeyboardInterrupt:
self.target.killall(self.binary_name)
raise
def update_output(self, context):
super(GeekbenchCmdline, self).update_output(context)
if not self.output:
return
for workload in self.workload:
scores = []
matches = re.findall(self.workloads[workload] + '(.+\d)', self.output)
for match in matches:
scores.append(int(re.search(r'\d+', match).group(0)))
if self.section == 4:
context.add_metric("OpenCL Score " + self.workloads[workload], scores[0])
elif self.section == 9:
context.add_metric("Vulkan Score " + self.workloads[workload], scores[0])
else:
context.add_metric("Single-Core Score " + self.workloads[workload], scores[0])
if not self.is_single_core:
context.add_metric("Multi-Core Score " + self.workloads[workload], scores[1])
def extract_results(self, context):
# Extract results on the target
super(GeekbenchCmdline, self).extract_results(context)
self.target.pull(self.target_result_json, self.host_result_json)
context.add_artifact('GeekbenchCmdline_results', self.host_result_json, kind='raw')
@once
def finalize(self, context):
if self.cleanup_assets:
self.target.remove(self.target_exec_directory)