xuchen-amd
148 lines
5.6 KiB
Python
148 lines
5.6 KiB
Python
##############################################################################bl
|
|
# MIT License
|
|
#
|
|
# Copyright (c) 2025 Advanced Micro Devices, Inc. All Rights Reserved.
|
|
#
|
|
# Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
# of this software and associated documentation files (the "Software"), to deal
|
|
# in the Software without restriction, including without limitation the rights
|
|
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
# copies of the Software, and to permit persons to whom the Software is
|
|
# furnished to do so, subject to the following conditions:
|
|
#
|
|
# The above copyright notice and this permission notice shall be included in all
|
|
# copies or substantial portions of the Software.
|
|
#
|
|
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
# SOFTWARE.
|
|
##############################################################################el
|
|
|
|
import copy
|
|
from pathlib import Path
|
|
|
|
from rocprof_compute_analyze.analysis_base import OmniAnalyze_Base
|
|
from rocprof_compute_tui.utils.tui_utils import (
|
|
get_top_kernels_and_dispatch_ids,
|
|
process_panels_to_dataframes,
|
|
)
|
|
from utils import file_io, parser, schema
|
|
from utils.kernel_name_shortener import kernel_name_shortener
|
|
from utils.logger import console_error, demarcate
|
|
|
|
|
|
class tui_analysis(OmniAnalyze_Base):
|
|
def __init__(self, args, supported_archs, path):
|
|
super().__init__(args, supported_archs)
|
|
self.path = str(path)
|
|
self.arch = None
|
|
self.raw_dfs = {}
|
|
self.kernel_dfs = {}
|
|
|
|
# -----------------------
|
|
# Required child methods
|
|
# -----------------------
|
|
@demarcate
|
|
def pre_processing(self):
|
|
self._profiling_config = file_io.load_profiling_config(self.path)
|
|
|
|
self._runs = self.initalize_runs()
|
|
|
|
if self.get_args().random_port:
|
|
console_error("--gui flag is required to enable --random-port")
|
|
|
|
self._runs[self.path].raw_pmc = file_io.create_df_pmc(
|
|
self.path,
|
|
self.get_args().nodes,
|
|
self.get_args().spatial_multiplexing,
|
|
self.get_args().kernel_verbose,
|
|
self.get_args().verbose,
|
|
self._profiling_config,
|
|
)
|
|
|
|
if self.get_args().spatial_multiplexing:
|
|
self._runs[self.path].raw_pmc = self.spatial_multiplex_merge_counters(
|
|
self._runs[self.path].raw_pmc
|
|
)
|
|
|
|
file_io.create_df_kernel_top_stats(
|
|
df_in=self._runs[self.path].raw_pmc,
|
|
raw_data_dir=self.path,
|
|
filter_gpu_ids=self._runs[self.path].filter_gpu_ids,
|
|
filter_dispatch_ids=self._runs[self.path].filter_dispatch_ids,
|
|
filter_nodes=self._runs[self.path].filter_nodes,
|
|
time_unit=self.get_args().time_unit,
|
|
max_stat_num=self.get_args().max_stat_num,
|
|
kernel_verbose=self.get_args().kernel_verbose,
|
|
)
|
|
|
|
kernel_name_shortener(
|
|
self._runs[self.path].raw_pmc, self.get_args().kernel_verbose
|
|
)
|
|
|
|
# 1. load top kernel
|
|
parser.load_kernel_top(
|
|
workload=self._runs[self.path], dir=self.path, args=self.get_args()
|
|
)
|
|
|
|
# 2. load table data for each kernel
|
|
self.raw_dfs.clear()
|
|
for idx in self._runs[self.path].raw_pmc.index:
|
|
kernel_df = self._runs[self.path].raw_pmc.loc[[idx]]
|
|
kernel_name = kernel_df.pmc_perf["Kernel_Name"].loc[idx]
|
|
this_dfs = copy.deepcopy(self._runs[self.path].dfs)
|
|
parser.eval_metric(
|
|
this_dfs,
|
|
self._runs[self.path].dfs_type,
|
|
self._runs[self.path].sys_info.iloc[0],
|
|
kernel_df,
|
|
self.get_args().debug,
|
|
self._profiling_config,
|
|
)
|
|
|
|
self.raw_dfs[kernel_name] = this_dfs
|
|
|
|
def initalize_runs(self, normalization_filter=None):
|
|
sysinfo_path = Path(self.path)
|
|
sys_info = file_io.load_sys_info(sysinfo_path.joinpath("sysinfo.csv"))
|
|
self.arch = sys_info.iloc[0]["gpu_arch"]
|
|
args = self.get_args()
|
|
self.generate_configs(
|
|
self.arch,
|
|
args.config_dir,
|
|
args.list_stats,
|
|
args.filter_metrics,
|
|
sys_info.iloc[0],
|
|
)
|
|
|
|
self.load_options(normalization_filter)
|
|
|
|
w = schema.Workload()
|
|
w.sys_info = file_io.load_sys_info(sysinfo_path.joinpath("sysinfo.csv"))
|
|
mspec = self.get_socs()[self.arch]._mspec
|
|
if args.specs_correction:
|
|
w.sys_info = parser.correct_sys_info(mspec, args.specs_correction)
|
|
w.avail_ips = w.sys_info["ip_blocks"].item().split("|")
|
|
w.dfs = copy.deepcopy(self._arch_configs[self.arch].dfs)
|
|
w.dfs_type = self._arch_configs[self.arch].dfs_type
|
|
self._runs[self.path] = w
|
|
|
|
return self._runs
|
|
|
|
@demarcate
|
|
def run_kernel_analysis(self):
|
|
self.kernel_dfs.clear()
|
|
for kernel_name, df in self.raw_dfs.items():
|
|
self.kernel_dfs[kernel_name] = process_panels_to_dataframes(
|
|
self.get_args(), df, self._arch_configs[self.arch], roof_plot=None
|
|
)
|
|
return self.kernel_dfs
|
|
|
|
@demarcate
|
|
def run_top_kernel(self):
|
|
return get_top_kernels_and_dispatch_ids(self._runs)
|