mindspore 2.6.0rc1__cp310-cp310-win_amd64.whl → 2.7.0rc1__cp310-cp310-win_amd64.whl

mindspore/profiler/parser/ascend_hccl_generator.py

@@ -1,271 +0,0 @@
-# Copyright 2023 Huawei Technologies Co., Ltd
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-# ============================================================================
-"""hccl analyse model"""
-import copy
-import csv
-import json
-import logging
-import os
-import stat
-import glob
-
-import numpy as np
-from mindspore.profiler.common.exceptions.exceptions import ProfilerIOException
-from mindspore.profiler.common.util import get_newest_file
-
-
-class AscendHCCLGenerator:
-    """Generate ascend hccl data from files."""
-
-    DEFAULT_MODEL_ID = 4294967295
-
-    def __init__(self, mindstudio_profiler_output, steptrace):
-        self.mindstudio_profiler_output = mindstudio_profiler_output
-        self.steptrace = steptrace
-        self.hccl_raw = []
-        self.hccl_data_df = np.dtype(
-            [('model_id', int), ('iteration_id', int), ('name', object), ('pid', int), ('tid', int), ('ts', float),
-             ('te', float), ('dur', float), ('ph', object),
-             ('task_type', object), ('link_info', object), ('transport_type', object), ('size', int), ('tag', object)])
-
-    @staticmethod
-    def _cost_analyse(iteration):
-        """analyse communication cost and wait cost"""
-        communication_cost = np.sum(iteration[iteration['name'] != 'Notify_Wait']['dur'])
-        wait_cost = np.sum(iteration[iteration['name'] == 'Notify_Wait']['dur'])
-        return communication_cost, wait_cost
-
-    @staticmethod
-    def _rdma_analyse(groupby_transport):
-        """rdma analyse"""
-        thread_groups = np.unique(groupby_transport['tid'])
-        thread_information = []
-        for thread_index in thread_groups:
-            groupby_thread = groupby_transport[groupby_transport['tid'] == thread_index]
-            rdma_communication_time = 0
-            rdma_communication_size = 0
-            rdma_communication_wait_time = 0
-            start_index = 0
-            end_index = groupby_thread.size - 2
-            while start_index < end_index:
-                first_task_type = groupby_thread[start_index]['task_type']
-                if first_task_type == 'RDMASend':
-                    second_index = start_index + 1
-                    third_index = start_index + 2
-                    second_task_type = groupby_thread[second_index]['task_type']
-                    third_task_type = groupby_thread[third_index]['task_type']
-                    if second_task_type == 'RDMASend' and third_task_type == 'Notify Wait':
-                        rdma_send_cost = groupby_thread[start_index]['dur']
-                        notify_record_cost = groupby_thread[second_index]['dur']
-                        notify_wait_cost = groupby_thread[third_index]['dur']
-                        rdma_communication_time += rdma_send_cost + notify_record_cost + notify_wait_cost
-                        rdma_communication_wait_time += notify_wait_cost
-                        rdma_communication_size += groupby_thread[start_index]['size'] + groupby_thread[second_index][
-                            'size']
-                        start_index += 2
-                start_index += 1
-            rdma_communication_wait_time = rdma_communication_wait_time / 1e3
-            rdma_communication_size = rdma_communication_size / 1e3
-            rdma_communication_time = rdma_communication_time / 1e3
-            rdma_bandwidth = rdma_communication_size / (rdma_communication_time / 1e3) \
-                if rdma_communication_size else 0
-            thread_information.append(
-                [rdma_communication_time, rdma_communication_size, rdma_bandwidth, rdma_communication_wait_time])
-        if len(thread_information) > 1:
-            thread_information = np.sum(thread_information, axis=0).tolist()
-
-        return thread_information
-
-    def parse(self):
-        """Analyse the original hccl data generator hccl data."""
-
-        raw_data = []
-        msprof_name = fr'{self.mindstudio_profiler_output}/msprof_*.json'
-        for msprof_file in get_newest_file(glob.glob(msprof_name)):
-            with open(msprof_file) as fr:
-                raw_data.extend(json.load(fr))
-
-        hccl_data = self._original_data_analyse(raw_data)
-
-        for model_id in np.unique(hccl_data['model_id']):
-            hccl_data_model = hccl_data[hccl_data['model_id'] == model_id]
-            for iteration_id in np.unique(hccl_data_model['iteration_id']):
-                hccl_data_model_iteration = hccl_data_model[hccl_data_model['iteration_id'] == iteration_id]
-
-                hccl_abstract_data = hccl_data_model_iteration[hccl_data_model_iteration['task_type'] == '']
-                hccl_detail_data = hccl_data_model_iteration[hccl_data_model_iteration['task_type'] != '']
-                hccl_abstract_data = np.sort(hccl_abstract_data, order='ts')
-                hccl_detail_data = np.sort(hccl_detail_data, order='ts')
-
-                tag = np.searchsorted(hccl_abstract_data['ts'], hccl_detail_data['ts'], side='right') - 1
-
-                hccl_detail_data['tag'] = [x[-1] for x in
-                                           np.char.split(hccl_abstract_data[tag]['name'].astype(str), sep='/')]
-
-                self.hccl_raw.append(self._iteration_analyse(hccl_detail_data, iteration_id))
-
-        self.hccl_raw = sorted(self.hccl_raw, key=lambda x: x[0])
-        if self.hccl_raw:
-            self.hccl_raw.append(copy.deepcopy(self.hccl_raw[-1]))
-            self.hccl_raw[-1][0] = '-'
-            for _, value in self.hccl_raw[-1][4].items():
-                value[0] = '-'
-
-    def write(self, hccl_raw_path):
-        """
-        Write the flops.csv and flops_summary.json
-
-        Args:
-            hccl_raw_path(str): hccl_raw.csv path.
-        """
-        try:
-            with os.fdopen(os.open(hccl_raw_path,
-                                   os.O_WRONLY | os.O_CREAT | os.O_TRUNC, stat.S_IWUSR | stat.S_IRUSR), 'w',
-                           newline='') as hccl_row:
-                writer = csv.writer(hccl_row)
-                writer.writerow(
-                    ['step_num', 'communication_cost', 'wait_cost', 'link_info', 'communication_operator_cost'])
-                for row in self.hccl_raw:
-                    row[3] = json.dumps(row[3])
-                    row[4] = json.dumps(row[4])
-                writer.writerows(self.hccl_raw)
-        except (IOError, OSError) as err:
-            logging.critical('Errot occurred when write aicore detail file: %s', err)
-            raise ProfilerIOException() from err
-        if os.path.exists(hccl_raw_path):
-            os.chmod(hccl_raw_path, stat.S_IREAD | stat.S_IWRITE)
-
-    def _original_data_analyse(self, original_data):
-        """analyse original data"""
-
-        groups_steptrace = {model_id: np.sort(self.steptrace[self.steptrace['Model ID'] == model_id],
-                                              order='Iteration ID')
-                            for model_id in np.unique(self.steptrace['Model ID'])}
-
-        hccl_pid = None
-        for row in original_data:
-            if row.get('ph') == 'M' and row.get('name') == 'process_name' and row.get('args', {}).get('name') == 'HCCL':
-                hccl_pid = row.get('pid')
-                break
-
-        target_data = []
-        for row in original_data:
-            model_id = row.get('args', {}).get('model id')
-            pid = row.get('pid')
-            if hccl_pid == pid and row.get('ph') == 'X' and model_id is not None:
-                name = row.get('name')
-                tid = row.get('tid')
-                ts = float(row.get('ts'))
-                dur = float(row.get('dur'))
-                te = ts + dur
-                ph = row.get('ph')
-                task_type = row.get('args', {}).get('task type', '')
-                src_rank = row.get('args', {}).get('src rank', 0)
-                dst_rank = row.get('args', {}).get('dst rank', 0)
-                if src_rank == int('0xffffffff', 16):
-                    src_rank = dst_rank
-                if dst_rank == int('0xffffffff', 16):
-                    dst_rank = src_rank
-                transport_type = row.get('args', {}).get('transport type', '')
-                if transport_type == 'LOCAL':
-                    src_rank, dst_rank = dst_rank, src_rank
-                link_info = str(src_rank) + '-' + str(dst_rank)
-                size = row.get('args', {}).get('size(Byte)', 0)
-                size = size if isinstance(size, int) else int(size, 16)
-                steptrace = groups_steptrace.get(model_id, None)
-                if steptrace is None:
-                    if self.DEFAULT_MODEL_ID != model_id:
-                        logging.warning('Could not find model: %s in hccl json, skip.', model_id)
-                    continue
-                tag = np.searchsorted(steptrace['Iteration End'], te * 1e-3, side='left')
-                iteration_id = steptrace[tag]['Iteration ID']
-                target_data.append(
-                    tuple([model_id, iteration_id, name, pid, tid,
-                           ts, te, dur, ph, task_type,
-                           link_info, transport_type, size, -1]))
-
-        hccl_data = np.array(target_data, dtype=self.hccl_data_df)
-
-        return hccl_data
-
-    def _iteration_analyse(self, hccl_detail_data, iteration):
-        """analyse data by iteration """
-        communication_cost, wait_cost = self._cost_analyse(hccl_detail_data)
-        link_info = self._link_info_analyse(hccl_detail_data)
-        communication_operator_cost = self._communication_operator_cost_analyse(hccl_detail_data, iteration)
-        return [iteration, communication_cost, wait_cost, link_info, communication_operator_cost]
-
-    def _link_info_analyse(self, hccl_detail_data):
-        """analyse link info data"""
-        groupby_iteration = hccl_detail_data[hccl_detail_data['task_type'] != 'Notify Record']
-        link_info_groups = np.unique(groupby_iteration['link_info'])
-        link_info_information = dict()
-        for link_info_index in link_info_groups:
-            groupby_link_info = groupby_iteration[groupby_iteration['link_info'] == link_info_index]
-            transport_groups = np.unique(groupby_iteration['transport_type'])
-            transport_information = dict()
-            for transport_index in transport_groups:
-                groupby_transport = groupby_link_info[groupby_link_info['transport_type'] == transport_index]
-                if transport_index == 'SDMA' and groupby_transport.size > 0:
-                    groupby_sdma = \
-                        groupby_transport[np.isin(groupby_transport['task_type'], ['Memcpy', 'Reduce Inline'])][
-                            ['dur', 'size']]
-                    sdma_communication_time = np.sum(groupby_sdma['dur']) * 1e-3
-                    sdma_communication_size = np.sum(groupby_sdma['size']) * 1e-3
-                    sdma_bandwidth = sdma_communication_size / sdma_communication_time * 1e-3 \
-                        if sdma_communication_time != 0 else 0
-                    transport_information['SDMA'] = [sdma_communication_time, sdma_communication_size, sdma_bandwidth]
-                elif transport_index == 'RDMA' and groupby_transport.size > 0:
-                    transport_information['RDMA'] = self._rdma_analyse(groupby_transport)
-            link_info_information[link_info_index] = transport_information
-        return link_info_information
-
-    def _communication_operator_cost_analyse(self, hccl_detail_data, iteration_index):
-        """analyse communication operator cost"""
-        groupby_iteration = hccl_detail_data[hccl_detail_data['task_type'] != 'Notify Record']
-        tag_groups = np.unique(groupby_iteration['tag'])
-        tag_information = dict()
-        for tag_index in tag_groups:
-            groupby_tag = groupby_iteration[groupby_iteration['tag'] == tag_index]
-            link_groups = np.unique(groupby_iteration['link_info'])
-            link_info_information = dict()
-            for link_info_index in link_groups:
-                groupby_link_info = groupby_tag[groupby_tag['link_info'] == link_info_index]
-                transport_groups = np.unique(groupby_link_info['transport_type'])
-                transport_information = dict()
-                for transport_index in transport_groups:
-                    groupby_transport = groupby_link_info[groupby_link_info['transport_type'] == transport_index]
-                    if transport_index == 'SDMA':
-                        groupby_sdma = \
-                            groupby_transport[np.isin(groupby_transport['task_type'], ['Memcpy', 'Reduce Inline'])][
-                                ['dur', 'size']]
-                        sdma_communication_time = np.sum(groupby_sdma['dur']) * 1e-3
-                        sdma_communication_size = np.sum(groupby_sdma['size']) * 1e-3
-                        sdma_bandwidth = sdma_communication_size / sdma_communication_time * 1e-3 \
-                            if sdma_communication_time != 0 else 0
-                        transport_information['SDMA'] = [
-                            sdma_communication_time, sdma_communication_size,
-                            sdma_bandwidth
-                        ]
-                    elif transport_index == 'RDMA':
-                        transport_information['RDMA'] = self._rdma_analyse(groupby_transport)
-                    link_info_information[link_info_index] = transport_information
-                communication_cost = np.sum(groupby_tag[groupby_tag['name'] != 'Notify_Wait']['dur'])
-                wait_cost = np.sum(groupby_tag[groupby_tag['name'] == 'Notify_Wait']['dur'])
-                tag_information[tag_index] = [
-                    str(iteration_index), communication_cost, wait_cost,
-                    link_info_information
-                ]
-        return tag_information

mindspore/profiler/parser/ascend_integrate_generator.py

@@ -1,42 +0,0 @@
-# Copyright 2024 Huawei Technologies Co., Ltd
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-# ============================================================================
-"""PROF csv data analyze module"""
-import os
-
-from mindspore.profiler.parser.ascend_analysis.file_manager import FileManager
-
-
-class AscendIntegrateGenerator:
-    """Generate ms profiler output csv"""
-
-    def __init__(self, mindstudio_profiler_output: str, ascend_ms_output_path: str):
-        self._mindstudio_profiler_output = mindstudio_profiler_output
-        self._ascend_ms_output_path = ascend_ms_output_path
-
-    def parse(self):
-        """Generate ms profiler output csv"""
-        self._generate_kernel_details()
-
-    def _generate_kernel_details(self):
-        """Generate kernel_details.csv"""
-        header_map = {
-            "Op Name": "Name", "OP Type": "Type", "Task Type": "Accelerator Core",
-            "Task Start Time(us)": "Start Time(us)", "Task Duration(us)": "Duration(us)",
-            "Task Wait Time(us)": "Wait Time(us)",
-        }
-        op_summary_file_list = FileManager.get_csv_file_list_by_start_name(self._mindstudio_profiler_output,
-                                                                           "op_summary")
-        kernel_details_file = os.path.join(self._ascend_ms_output_path, "kernel_details.csv")
-        FileManager.combine_csv_file(op_summary_file_list, kernel_details_file, header_map)

mindspore/profiler/parser/ascend_memory_generator.py

@@ -1,185 +0,0 @@
-# Copyright 2024 Huawei Technologies Co., Ltd
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-# ============================================================================
-"""memory data analyze module"""
-import os
-
-from decimal import Decimal
-
-from mindspore.profiler.parser.ascend_analysis.file_manager import FileManager
-
-
-class AscendMemoryGenerator:
-    """Parsing the memory data of the ascend device"""
-    GE_MEMORY_RECORD_HEADERS = ["Device id", "Component", "Timestamp(us)", "Total Allocated(KB)",
-                                "Total Reserved(KB)", "Device"]
-    MS_MEMORY_RECORD_HEADERS = ["Timestamp(ns)", "Total Allocated(Byte)", "Total Reserved(Byte)", "Total Active(Byte)"]
-    TARGET_MEMORY_RECORD_HEADERS = ["Component", "Timestamp(us)", "Total Allocated(MB)", "Total Reserved(MB)",
-                                    "Total Active(MB)", "Device Type"]
-
-    def __init__(self, source_path: str, rank_id: int, mindstudio_profiler_output: str, ascend_ms_output_path: str):
-        self._source_path = source_path
-        self._rank_id = rank_id
-        self._mindstudio_profiler_output = mindstudio_profiler_output
-        self._ascend_ms_output_path = ascend_ms_output_path
-        self._ge_memory_record = []
-        self._ms_memory_record = []
-
-    def parse(self):
-        """Parse all memory data"""
-        self.copy_npu_module_mem_csv()
-        self.parse_memory_record()
-
-    def copy_npu_module_mem_csv(self):
-        """Generate npu_module_mem.csv"""
-        npu_module_mem_file_list = FileManager.get_csv_file_list_by_start_name(self._mindstudio_profiler_output,
-                                                                               "npu_module_mem")
-        target_file_path = os.path.join(self._ascend_ms_output_path, "npu_module_mem.csv")
-        FileManager.combine_csv_file(npu_module_mem_file_list, target_file_path)
-
-    def parse_memory_record(self):
-        """Generate memory_record.csv"""
-        self._parse_ge_memory_record()
-        self._parse_ms_memory_record()
-        combined_memory_data = self._combine_ge_ms_memory_record()
-        target_file_path = os.path.join(self._ascend_ms_output_path, "memory_record.csv")
-        FileManager.create_csv_file(target_file_path, combined_memory_data, self.TARGET_MEMORY_RECORD_HEADERS)
-
-    def _parse_ge_memory_record(self):
-        """Parse ge memory record data"""
-        memory_record_file_list = FileManager.get_csv_file_list_by_start_name(self._mindstudio_profiler_output,
-                                                                              "memory_record")
-        for file in memory_record_file_list:
-            data = FileManager.read_csv_file(file)
-            if len(data) > 1:
-                self._ge_memory_record.extend(data[1:])
-
-    def _parse_ms_memory_record(self):
-        """Parse mindspore memory record data"""
-        memory_record_file = os.path.join(self._source_path, f"cpu_ms_memory_record_{self._rank_id}.txt")
-        data = FileManager.read_txt_file(memory_record_file)
-        if len(data) > 1:
-            self._ms_memory_record.extend(data[1:])
-
-    def _get_app_reserved_memory(self) -> list:
-        """Get the reserved memory of the application from npu_mem.csv"""
-        npu_module_mem_file_list = FileManager.get_csv_file_list_by_start_name(self._mindstudio_profiler_output,
-                                                                               "npu_mem")
-        app_mems = []
-        for file in npu_module_mem_file_list:
-            md_mems = FileManager.read_csv_file(file)
-            for mem in md_mems:
-                if mem[1] == "APP":
-                    app_mems.append(MemoryRecordBean([mem[1], mem[-1].rstrip('\t'), 0.0, float(mem[4]), 0.0,
-                                                      f"NPU:{self._rank_id}"]).row)
-
-        return app_mems
-
-    def _combine_ge_ms_memory_record(self) -> list:
-        """Combine ge and mindspore memory record data"""
-        memory_records = []
-        for ge_memory in self._ge_memory_record:
-            memory_record = dict(zip(self.GE_MEMORY_RECORD_HEADERS, ge_memory))
-            memory_records.append(
-                MemoryRecordBean([memory_record.get("Component", "GE"), memory_record.get("Timestamp(us)"),
-                                  memory_record.get("Total Allocated(KB)", 0),
-                                  memory_record.get("Total Reserved(KB)", 0),
-                                  0, memory_record.get("Device")]))
-        for ms_memory in self._ms_memory_record:
-            memory_record = dict(zip(self.MS_MEMORY_RECORD_HEADERS, ms_memory))
-            memory_records.append(
-                MemoryRecordBean(["MindSpore", Decimal(memory_record.get("Timestamp(ns)", 0)) / 1000,
-                                  float(memory_record.get("Total Allocated(Byte)", 0)) / 1024,
-                                  float(memory_record.get("Total Reserved(Byte)", 0)) / 1024,
-                                  float(memory_record.get("Total Active(Byte)", 0)) / 1024,
-                                  f"NPU:{self._rank_id}"]))
-        memory_records.sort(key=lambda x: x.time_us)
-        last_ge_memory, last_ms_memory = MemoryRecordBean([0] * 6), MemoryRecordBean([0] * 6)
-        result_data = []
-        for memory_record in memory_records:
-            result_data.append(memory_record.row)
-            last_memory = last_ms_memory if memory_record.is_ge_component() else last_ge_memory
-            combined_mem = MemoryRecordBean(["MindSpore+GE", memory_record.time_us,
-                                             memory_record.total_allocated_kb + last_memory.total_allocated_kb,
-                                             memory_record.total_reserved_kb + last_memory.total_reserved_kb,
-                                             memory_record.total_active_kb + last_memory.total_active_kb,
-                                             f"NPU:{self._rank_id}"])
-            result_data.append(combined_mem.row)
-            if memory_record.is_ge_component():
-                last_ge_memory = memory_record
-            else:
-                last_ms_memory = memory_record
-        return result_data + self._get_app_reserved_memory()
-
-
-class MemoryRecordBean:
-    """Memory Record Struct"""
-    KB_TO_MB = 1000
-    HEADERS = ["Component", "Timestamp(us)", "Total Allocated(KB)", "Total Reserved(KB)",
-               "Total Active(KB)", "Device Type"]
-
-    def __init__(self, data: list):
-        self._data = dict(zip(self.HEADERS, data))
-
-    @property
-    def row(self) -> list:
-        return [self.component, self.time_us_str, self.total_allocated_mb, self.total_reserved_mb,
-                self.total_active_mb, self.device_type]
-
-    @property
-    def component(self) -> str:
-        return self._data.get("Component", "")
-
-    @property
-    def time_us_str(self) -> str:
-        ts_us = self._data.get("Timestamp(us)", 0)
-        return str(ts_us)
-
-    @property
-    def total_allocated_mb(self) -> float:
-        return float(self._data.get("Total Allocated(KB)", 0)) / self.KB_TO_MB
-
-    @property
-    def total_reserved_mb(self) -> float:
-        return float(self._data.get("Total Reserved(KB)", 0)) / self.KB_TO_MB
-
-    @property
-    def total_active_mb(self) -> float:
-        return float(self._data.get("Total Active(KB)", 0)) / self.KB_TO_MB
-
-    @property
-    def device_type(self) -> float:
-        return self._data.get("Device Type", "")
-
-    @property
-    def total_allocated_kb(self) -> float:
-        return float(self._data.get("Total Allocated(KB)", 0))
-
-    @property
-    def total_reserved_kb(self) -> float:
-        return float(self._data.get("Total Reserved(KB)", 0))
-
-    @property
-    def total_active_kb(self) -> float:
-        return float(self._data.get("Total Active(KB)", 0))
-
-    @property
-    def time_us(self) -> Decimal:
-        return Decimal(self._data.get("Timestamp(us)", 0))
-
-    def is_ge_component(self):
-        """
-        Determine if it is GE
-        """
-        return self.component == "GE"