o NeyE@sddlmZddlmZddlmZddlZddlZddlZddl m Z gdZ gdZ e d d dd d edd dZ edddZe d d dd d ddZe d d dd d d ddZe d d dd d d!ddZe d d dd d e   d"ddZed#ddZdS)$)print_function)core)signature_safe_contextmanagerN) deprecated) cuda_profilerreset_profilerprofilerstart_profiler stop_profiler)ZgpustarttimestampZgpuendtimestampZ gridsize3dZthreadblocksizeZstreamidzenableonstart 0Zconckerneltracez2.3.0zpaddle.profiler.ProfilerzIPlease use new profiler tool, this profiler tool is no longer maintained.)ZsinceZ update_tolevelreasoncCstd)a\ API cuda_profiler has been abandoned. If you have relevant requirements, you can use `paddle.utils.profiler.start_profiler` and `paddle.utils.profiler.stop_profiler`. The relevant reference documents are as follows: aBAPI cuda_profiler has been abandoned. If you have relevant requirements, you can use `paddle.utils.profiler.start_profiler` and `paddle.utils.profiler.stop_profiler`. The relevant reference documents are as follows: ) RuntimeError) output_fileZ output_modeconfigrED:\Projects\ConvertPro\env\Lib\site-packages\paddle/fluid/profiler.pyr)src csV|st}t|t|zdVWt|tdSt|tw)a The NPU profiler. This fuctions is used to profile NPU program by NPU runtime application programming interface. The profiling result will be written into `output_file`. The users can set set the NPU profiling config by `config` argument. After getting the profiling result file, users can use `tools provided by Ascend `_ to load this output file to visualize results. Args: output_file (str) : The output file name, the result will be written into this file. It should be absolute path. config (list, optional) : NPU profile config. For more details, please refer to `User Guide `_ . Examples: .. code-block:: python import paddle.fluid as fluid import paddle.fluid.profiler as profiler import numpy as np epoc = 8 dshape = [4, 3, 28, 28] data = fluid.data(name='data', shape=[None, 3, 28, 28], dtype='float32') conv = fluid.layers.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1]) place = fluid.NPUPlace(0) exe = fluid.Executor(place) exe.run(fluid.default_startup_program()) output_file = 'npu.txt' with profiler.npu_profiler(output_file) as npu_prof: for i in range(epoc): input = np.random.random(dshape).astype('float32') exe.run(fluid.default_main_program(), feed={'data': input}) # then use NPU profiler tools to load this output file # to visualize results. N)rZnpu_prof_create_configZ npu_prof_initZnpu_prof_startZ npu_prof_stopZnpu_prof_finalize)rrrrr npu_profiler=s-     rcCs tdS)a! Clear the previous time record. It works for `fluid.profiler.start_profiler`, `fluid.profiler.stop_profiler`, and `fluid.profiler.profiler`. Examples: .. code-block:: python # required: gpu import paddle.fluid as fluid import paddle.fluid.profiler as profiler with profiler.profiler('CPU', 'total', '/tmp/profile'): for iter in range(10): if iter == 2: profiler.reset_profiler() # ... N)rrrrrrrxs rDefaultcCstrdS|dvrtd|dkrtjj}n |dkr tjj}ntjj}|dvr,td|dkr5tjj}n |d kr>tjj }ntjj }t |t |dS) a. Enable the profiler. Uers can use `fluid.profiler.start_profiler` and `fluid.profiler.stop_profiler` to profile, which is equal to the usage of `fluid.profiler.profiler` interface. Args: state (str) : The profiling state, which should be one of 'CPU', 'GPU' or 'All'. 'CPU' means only profiling CPU; 'GPU' means profiling both CPU and GPU; 'All' means profiling both CPU and GPU, and generates timeline as well. tracer_option (str, optional) : tracer_option can be one of ['Default', 'OpDetail', 'AllOpDetail'], it can control the profile level and print the different level profile result. `Default` option print the different Op type profiling result and the `OpDetail` option print the detail profiling result of different op types such as compute and data transform, `AllOpDetail` option print the detail profiling result of different op name same as `OpDetail`. Raises: ValueError: If `state` is not in ['CPU', 'GPU', 'All'] or `tracer_option` is not in ['Default', 'OpDetail', 'AllOpDetail']. Examples: .. code-block:: python # required: gpu import paddle.fluid as fluid import paddle.fluid.profiler as profiler profiler.start_profiler('GPU') for iter in range(10): if iter == 2: profiler.reset_profiler() # except each iteration profiler.stop_profiler('total', '/tmp/profile') profiler.start_profiler('GPU', "OpDetail") for iter in range(10): if iter == 2: profiler.reset_profiler() # except each iteration profiler.stop_profiler('total', '/tmp/profile') N)CPUGPUZAllz*The state must be 'CPU' or 'GPU' or 'All'.rr)rOpDetailZ AllOpDetailz;tracer option must be 'Default', 'OpDetail', 'AllOpDetail'.rr) ris_profiler_enabled ValueErrorZ ProfilerStateZkCUDAZkCPUZkAllZ TracerOptionkDefaultZ kOpDetailZ kAllOpDetailZset_tracer_optionZenable_profiler)state tracer_optionZ prof_stateZprof_tracer_optionrrrr s(1     r /tmp/profilecCsjtsdS|dur dn|}|dvrtdtjjtjjtjjtjjtjjtjj d}t |||dS)a Stop the profiler. Uers can use `fluid.profiler.start_profiler` and `fluid.profiler.stop_profiler` to profile, which is equal to the usage of `fluid.profiler.profiler` interface. Args: sorted_key (str, optional) : The order of profiling results, which should be one of None, 'calls', 'total', 'max', 'min' or 'ave'. Default is None, means the profiling results will be printed in the order of first end time of events. The `calls` means sorting by the number of calls. The `total` means sorting by the total execution time. The `max` means sorting by the maximum execution time. The `min` means sorting by the minimum execution time. The `ave` means sorting by the average execution time. and write it into `profile_path`. The default profile_path is `/tmp/profile`. profile_path (str, optional) : If state == 'All', it will generate timeline, Raises: ValueError: If `sorted_key` is not in ['calls', 'total', 'max', 'min', 'ave']. Examples: .. code-block:: python # required: gpu import paddle.fluid as fluid import paddle.fluid.profiler as profiler profiler.start_profiler('GPU') for iter in range(10): if iter == 2: profiler.reset_profiler() # except each iteration profiler.stop_profiler('total', '/tmp/profile') Ndefault)rZcallstotalmaxminZavezJThe sorted_key must be None or in 'calls', 'total', 'max', 'min' and 'ave') rrrZEventSortingKeyrZkCallsZkTotalZkMaxZkMinZkAveZdisable_profiler) sorted_key profile_pathZkey_maprrrr s, r c cs0t||z dVWt||dSt||w)ab The profiler interface. This profiler can be used to profile both CPU and GPU program. Args: state (str) : The profiling state, which should be one of 'CPU', 'GPU' or 'All'. 'CPU' means only profiling CPU; 'GPU' means profiling both CPU and GPU; 'All' means profiling both CPU and GPU, and generates timeline as well. sorted_key (str, optional) : The order of profiling results, which should be one of None, 'calls', 'total', 'max', 'min' or 'ave'. Default is None, means the profiling results will be printed in the order of first end time of events. The `calls` means sorting by the number of calls. The `total` means sorting by the total execution time. The `max` means sorting by the maximum execution time. The `min` means sorting by the minimum execution time. The `ave` means sorting by the average execution time. profile_path (str, optional) : If state == 'All', it will generate timeline, and write it into `profile_path`. The default profile_path is `/tmp/profile`. tracer_option (str, optional) : tracer_option can be one of ['Default', 'OpDetail', 'AllOpDetail'], it can control the profile level and print the different level profile result. `Default` option print the different Op type profiling result and the `OpDetail` option print the detail profiling result of different op types such as compute and data transform, `AllOpDetail` option print the detail profiling result of different op name same as `OpDetail`. Raises: ValueError: If `state` is not in ['CPU', 'GPU', 'All']. If `sorted_key` is not in ['calls', 'total', 'max', 'min', 'ave']. Examples: .. code-block:: python # required: gpu import paddle.fluid as fluid import paddle.fluid.profiler as profiler import numpy as np import paddle paddle.enable_static() epoc = 8 dshape = [4, 3, 28, 28] data = fluid.data(name='data', shape=[None, 3, 28, 28], dtype='float32') conv = fluid.layers.conv2d(data, 20, 3, stride=[1, 1], padding=[1, 1]) place = fluid.CPUPlace() exe = fluid.Executor(place) exe.run(fluid.default_startup_program()) with profiler.profiler('CPU', 'total', '/tmp/profile', 'Default') as prof: for i in range(epoc): input = np.random.random(dshape).astype('float32') exe.run(fluid.default_main_program(), feed={'data': input}) Examples Results: .. code-block:: text #### Examples Results #### #### 1) sorted_key = 'total', 'calls', 'max', 'min', 'ave' #### # The only difference in 5 sorted_key results is the following sentence: # "Sorted by number of xxx in descending order in the same thread." # The reason is that in this example, above 5 columns are already sorted. -------------------------> Profiling Report <------------------------- Place: CPU Time unit: ms Sorted by total time in descending order in the same thread #Sorted by number of calls in descending order in the same thread #Sorted by number of max in descending order in the same thread #Sorted by number of min in descending order in the same thread #Sorted by number of avg in descending order in the same thread Event Calls Total Min. Max. Ave. Ratio. thread0::conv2d 8 129.406 0.304303 127.076 16.1758 0.983319 thread0::elementwise_add 8 2.11865 0.193486 0.525592 0.264832 0.016099 thread0::feed 8 0.076649 0.006834 0.024616 0.00958112 0.000582432 #### 2) sorted_key = None #### # Since the profiling results are printed in the order of first end time of Ops, # the printed order is feed->conv2d->elementwise_add -------------------------> Profiling Report <------------------------- Place: CPU Time unit: ms Sorted by event first end time in descending order in the same thread Event Calls Total Min. Max. Ave. Ratio. thread0::feed 8 0.077419 0.006608 0.023349 0.00967738 0.00775934 thread0::conv2d 8 7.93456 0.291385 5.63342 0.99182 0.795243 thread0::elementwise_add 8 1.96555 0.191884 0.518004 0.245693 0.196998 N)r r )rr"r#rrrrr s  gr Tccsz7||krtt||krtt|dVW||kr%t||kr5t|r7tdSdSdS||krAt||krPt|rQtwww)a  A range profiler interface (not public yet). Examples: .. code-block:: python model = Model() for i in range(max_iter): paddle.fluid.profiler._nvprof_range(i, 10, 20): out = model(in) N) rZ nvprof_startZnvprof_enable_record_eventZnvprof_nvtx_pushstrZnvprof_nvtx_popZ nvprof_stopsysexit)Ziter_idstartendZexit_after_profrrr _nvprof_ranges0  r))NN)N)r)Nr)Nrr)T) __future__rrwrapped_decoratorrossixr%Zpaddle.utils.deprecatedr__all__Z NVPROF_CONFIGrrrr r r r)rrrrsj      :  D 9g