Speed

[TheodoreEhrenborg]

I need to make the algorithms faster so that I can get statistically significant data. I think I'll split up aux_stochastic and aux_abridged_game among their users. I'll focus on making aux_stochastic faster with Cython and numpy Remember to set the array's length beforehand; appending takes time. See https://stackoverflow.com/questions/50171246/how-to-define-a-list-in-cython

Here's the profile data on running 5 copies of neural_nash against each other. Using profile slowed the program down by a factor of 10.

Mon Feb 25 17:16:25:mercantilism jtae$ python Python 2.7.10 (default, Oct 6 2017, 22:29:07) [GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.31)] on darwin Type "help", "copyright", "credits" or "license" for more information.

import algorithms algorithms.Neural_Evolver.evolve(generations = 10**9, trials_per_generation = 1000, population = 100) 2019-02-25 17:17:01.102140: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA ^CTraceback (most recent call last): File "", line 1, in File "algorithms.py", line 256, in evolve g = Game( playing_players, copy.copy(TOKENS) ) File "algorithms.py", line 614, in init this_move = player_list[0].choose_token( copy.deepcopy(self.tokens), copy.deepcopy(data), self.name ) File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/copy.py", line 163, in deepcopy y = copier(x, memo) File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/copy.py", line 230, in _deepcopy_list y.append(deepcopy(a, memo)) KeyboardInterrupt import random random.choice <bound method Random.choice of <random.Random object at 0x7fbad19abc20>> quit() Tue Feb 26 15:36:33:mercantilism jtae$ python Python 2.7.10 (default, Oct 6 2017, 22:29:07) [GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.31)] on darwin Type "help", "copyright", "credits" or "license" for more information. import profile temp = inspect.getmembers( algorithms, inspect.isfunction) Traceback (most recent call last): File "", line 1, in NameError: name 'inspect' is not defined import inspect temp = inspect.getmembers( algorithms, inspect.isfunction) Traceback (most recent call last): File "", line 1, in NameError: name 'algorithms' is not defined import algorithms temp = inspect.getmembers( algorithms, inspect.isfunction) temp [('aux_abridged_game', <function aux_abridged_game at 0x114602d70>), ('aux_exp', <function aux_exp at 0x1145cbc80>), ('aux_list_total', <function aux_list_total at 0x1146055f0>), ('aux_power', <function aux_power at 0x1145cbde8>), ('aux_stochastic', <function aux_stochastic at 0x114605938>), ('best_human_strategy', <function best_human_strategy at 0x1145f9e60>), ('exp_2', <function exp_2 at 0x1145cbcf8>), ('neural_evolve', <function neural_evolve at 0x1146022a8>), ('neural_nash', <function neural_nash at 0x114602758>), ('play_highest', <function play_highest at 0x1145fd320>), ('power_1', <function power_1 at 0x1145f9de8>), ('power_2', <function power_2 at 0x1145d4848>), ('power_3', <function power_3 at 0x1145d4e60>), ('quick_evolve', <function quick_evolve at 0x114602230>), ('round_diff', <function round_diff at 0x114605758>), ('round_points', <function round_points at 0x114605668>), ('round_winner', <function round_winner at 0x114605848>), ('total_diff', <function total_diff at 0x1146057d0>), ('total_points', <function total_points at 0x1146056e0>), ('total_winner', <function total_winner at 0x1146058c0>), ('uniform', <function uniform at 0x1145cbc08>)] temp[7] ('neural_evolve', <function neural_evolve at 0x1146022a8>) temp[8] ('neural_nash', <function neural_nash at 0x114602758>) x = temp[8] x ('neural_nash', <function neural_nash at 0x114602758>) import api 2019-02-26 15:37:39.017657: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA w = api.Wrapper(x) w Wrapper(('neural_nash', <function neural_nash at 0x114602758>)) api.API.TOKENS Traceback (most recent call last): File "", line 1, in AttributeError: class API has no attribute 'TOKENS' tokens = range(1,15) tokens = range(1,16) tokens [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15] import copy algorithm_tuple = (w,w,w,w,w) g = Game( copy.copy(algorithm_tuple), copy.copy(tokens) ) Traceback (most recent call last): File "", line 1, in NameError: name 'Game' is not defined g = api.Game( copy.copy(algorithm_tuple), copy.copy(tokens) ) help(g)

g.get_results() Traceback (most recent call last): File "", line 1, in TypeError: get_results() takes no arguments (1 given) help(g)

profile.run( g2 = api.Game( copy.copy(algorithm_tuple), copy.copy(tokens) ) ) ^CTraceback (most recent call last): File "", line 1, in File "api.py", line 482, in init this_move = Game.Neural_Nash_Instance.actually_choose_token( copy.deepcopy(self.tokens), copy.deepcopy(data), self.name ) File "algorithms.py", line 549, in actually_choose_token return aux_stochastic(tokens, data, game_name, self) File "algorithms.py", line 839, in aux_stochastic my_utilities[l].append( temp_utilities[i] ) KeyboardInterrupt profile.run( 2+2 ) Traceback (most recent call last): File "", line 1, in File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/profile.py", line 61, in run prof = prof.run(statement) File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/profile.py", line 438, in run return self.runctx(cmd, dict, dict) File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/profile.py", line 444, in runctx exec cmd in globals, locals TypeError: exec: arg 1 must be a string, file, or code object profile.run( print 2+2 ) File "", line 1 profile.run( print 2+2 ) ^ SyntaxError: invalid syntax profile.run( "print 2+2" ) 4 3 function calls in 0.003 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function) 1 0.003 0.003 0.003 0.003 :0(setprofile) 1 0.000 0.000 0.000 0.000 :1() 1 0.000 0.000 0.003 0.003 profile:0(print 2+2) 0 0.000 0.000 profile:0(profiler)

profile.run( "print 2**100" ) 1267650600228229401496703205376 3 function calls in 0.000 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function) 1 0.000 0.000 0.000 0.000 :0(setprofile) 1 0.000 0.000 0.000 0.000 :1() 1 0.000 0.000 0.000 0.000 profile:0(print 2**100) 0 0.000 0.000 profile:0(profiler)

profile.run( "print 2+2" ) 4 3 function calls in 0.000 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function) 1 0.000 0.000 0.000 0.000 :0(setprofile) 1 0.000 0.000 0.000 0.000 :1() 1 0.000 0.000 0.000 0.000 profile:0(print 2+2) 0 0.000 0.000 profile:0(profiler)

profile.run( "g2 = api.Game( copy.copy(algorithm_tuple), copy.copy(tokens) )" ) 32030683 function calls (31381333 primitive calls) in 662.242 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function) 32400 0.574 0.000 0.574 0.000 :0(Flatten) 32400 0.288 0.000 0.288 0.000 :0(TFE_Py_TapeSetRestartOnThread) 32400 0.299 0.000 0.299 0.000 :0(TFE_Py_TapeSetStopOnThread) 32400 0.293 0.000 0.293 0.000 :0(TF_DeleteStatus) 32400 0.307 0.000 0.307 0.000 :0(TF_GetCode) 32400 0.350 0.000 0.350 0.000 :0(TF_NewStatus) 32400 14.968 0.000 14.968 0.000 :0(TF_SessionRunCallable) 97200 0.801 0.000 0.801 0.000 :0(new) 32400 0.238 0.000 0.238 0.000 :0(add) 97200 1.173 0.000 4.316 0.000 :0(any) 11436865 121.231 0.000 121.231 0.000 :0(append) 32400 0.609 0.000 0.609 0.000 :0(arange) 97200 2.078 0.000 2.078 0.000 :0(array) 69 0.002 0.000 0.002 0.000 :0(asctime) 69 0.020 0.000 0.020 0.000 :0(close) 732 0.008 0.000 0.008 0.000 :0(count) 471636 3.549 0.000 3.549 0.000 :0(get) 778338 6.410 0.000 6.410 0.000 :0(getattr) 1134000 10.384 0.000 20.577 0.000 :0(hasattr) 2380 0.018 0.000 0.018 0.000 :0(id) 26400 0.332 0.000 0.332 0.000 :0(index) 1491900 14.239 0.000 20.644 0.000 :0(isinstance) 1626069 15.276 0.000 15.276 0.000 :0(len) 27132 0.375 0.000 0.375 0.000 :0(max) 58800 0.569 0.000 0.569 0.000 :0(min) 69 0.006 0.000 0.006 0.000 :0(open) 226830 1.803 0.000 1.803 0.000 :0(pop) 1248721 8.620 0.000 8.620 0.000 :0(random) 1515336 17.637 0.000 17.637 0.000 :0(range) 32400 0.252 0.000 0.252 0.000 :0(remove) 5 0.000 0.000 0.000 0.000 :0(repr) 32400 0.275 0.000 0.275 0.000 :0(setdefault) 1 0.000 0.000 0.000 0.000 :0(setprofile) 3 0.000 0.000 0.000 0.000 :0(time) 32400 0.289 0.000 0.289 0.000 :0(tolist) 69 0.001 0.000 0.001 0.000 :0(write) 32400 0.497 0.000 0.497 0.000 :0(zeros) 97200 1.092 0.000 1.092 0.000 :0(zip) 1 0.000 0.000 662.241 662.241 :1() 97200 1.472 0.000 2.273 0.000 :8(new) 32400 0.259 0.000 0.259 0.000 _weakrefset.py:16(init) 32400 0.506 0.000 0.744 0.000 _weakrefset.py:20(enter) 32400 0.988 0.000 1.531 0.000 _weakrefset.py:26(exit) 32400 0.501 0.000 0.501 0.000 _weakrefset.py:36(init) 32400 0.291 0.000 0.291 0.000 _weakrefset.py:52(_commit_removals) 680400 6.134 0.000 8.669 0.000 _weakrefset.py:58(iter) 195876 1.859 0.000 1.859 0.000 _weakrefset.py:70(contains) 97938 3.753 0.000 6.405 0.000 abc.py:128(instancecheck) 33132 17.208 0.001 330.318 0.010 algorithms.py:512(aux_evaluate_position) 30 0.040 0.001 662.046 22.068 algorithms.py:548(actually_choose_token) 33132 17.061 0.001 368.784 0.011 algorithms.py:697(aux_abridged_game) 198000 3.940 0.000 3.940 0.000 algorithms.py:779(aux_list_total) 30 148.376 4.946 662.006 22.067 algorithms.py:796(aux_stochastic) 1 0.011 0.011 662.241 662.241 api.py:449(init) 65 0.000 0.000 0.000 0.000 api.py:569(get_name) 5 0.000 0.000 0.000 0.000 api.py:573(get_tuple) 5 0.000 0.000 0.000 0.000 api.py:575(repr) 35 0.001 0.000 0.001 0.000 api.py:577(str) 32400 0.550 0.000 0.844 0.000 backend.py:2938(_call_fetch_callbacks) 32400 8.261 0.000 97.749 0.003 backend.py:2943(call) 32400 2.509 0.000 88.647 0.003 backend.py:359(learning_phase) 32400 2.205 0.000 53.857 0.002 backend.py:442(get_session) 64800 0.949 0.000 1.475 0.000 backend.py:586(is_sparse) 32400 1.616 0.000 4.828 0.000 backend.py:711(_get_variables) 32400 11.397 0.000 32.048 0.001 backend.py:720(_initialize_variables) 32400 0.451 0.000 0.702 0.000 base.py:702(isspmatrix) 194400 1.441 0.000 1.441 0.000 base_layer.py:236(trainable_weights) 194400 1.426 0.000 1.426 0.000 base_layer.py:240(non_trainable_weights) 194400 4.534 0.000 7.401 0.000 base_layer.py:255(weights) 194400 2.791 0.000 10.192 0.000 base_layer.py:264(variables) 32400 0.496 0.000 0.847 0.000 c_api_util.py:32(init) 32400 0.494 0.000 0.787 0.000 c_api_util.py:35(del) 65538 2.173 0.000 17.469 0.000 collections.py:446(init) 246846 1.682 0.000 1.682 0.000 collections.py:467(missing) 65538 5.740 0.000 14.847 0.000 collections.py:516(update) 97200 2.522 0.000 5.591 0.000 context.py:122(push) 97200 1.463 0.000 2.189 0.000 context.py:138(pop) 194400 2.011 0.000 2.011 0.000 context.py:343(_mode) 356400 2.884 0.000 2.884 0.000 context.py:364(executing_eagerly) 194400 1.378 0.000 1.378 0.000 context.py:606(context_switches) 648000 4.569 0.000 4.569 0.000 context.py:628(context) 324000 7.111 0.000 12.050 0.000 context.py:655(executing_eagerly) 97200 2.213 0.000 5.158 0.000 context.py:671(graph_mode) 388800 2.813 0.000 2.813 0.000 contextlib.py:12(init) 388800/64800 6.285 0.000 69.946 0.001 contextlib.py:15(enter) 388800/64800 8.077 0.000 28.599 0.000 contextlib.py:21(exit) 388800 6.713 0.000 9.526 0.000 contextlib.py:82(helper) 1 0.000 0.000 0.000 0.000 copy.py:101(_copy_immutable) 31 0.000 0.000 0.000 0.000 copy.py:113(_copy_with_constructor) 1110/60 0.051 0.000 0.144 0.002 copy.py:145(deepcopy) 490 0.003 0.000 0.003 0.000 copy.py:198(_deepcopy_atomic) 360/60 0.024 0.000 0.137 0.002 copy.py:226(_deepcopy_list) 850 0.018 0.000 0.029 0.000 copy.py:267(_keep_alive) 32 0.001 0.000 0.001 0.000 copy.py:66(copy) 32400 0.764 0.000 3.987 0.000 data_utils.py:97(is_generator_or_sequence) 64800 0.508 0.000 0.508 0.000 dtypes.py:125(as_numpy_dtype) 64800 1.038 0.000 1.518 0.000 dtypes.py:661(as_dtype) 32400 0.603 0.000 1.449 0.000 errors_impl.py:518(enter) 32400 1.130 0.000 2.224 0.000 errors_impl.py:522(exit) 32400 1.708 0.000 2.238 0.000 generic_utils.py:457(make_batches) 32400 2.623 0.000 3.933 0.000 generic_utils.py:472(slice_arrays) 32400 0.475 0.000 0.719 0.000 inspect.py:164(isgenerator) 194400 5.300 0.000 21.221 0.000 layer_utils.py:25(is_layer) 32400 3.268 0.000 25.922 0.001 layer_utils.py:40(filter_empty_layer_containers) 32400 0.807 0.000 1.281 0.000 network.py:441(uses_learning_phase) 32400 2.170 0.000 31.267 0.001 network.py:446(stateful) 32400 0.587 0.000 26.510 0.001 network.py:514(layers) 64800 1.013 0.000 2.234 0.000 numeric.py:469(asarray) 162000 2.284 0.000 3.844 0.000 ops.py:156(is_dense_tensor_like) 32400 0.251 0.000 0.251 0.000 ops.py:3022(_get_control_flow_context) 64800 0.470 0.000 0.470 0.000 ops.py:3030(_set_control_flow_context) 162000 1.164 0.000 1.164 0.000 ops.py:3187(building_function) 64800 0.478 0.000 0.478 0.000 ops.py:341(dtype) 97200 1.619 0.000 4.096 0.000 ops.py:3706(as_default) 64800 0.990 0.000 1.478 0.000 ops.py:3895(_name_stack) 64800 0.523 0.000 0.523 0.000 ops.py:3902(_name_stack) 64800 1.344 0.000 2.605 0.000 ops.py:3907(name_scope) 32400 0.511 0.000 1.252 0.000 ops.py:4074(get_name_scope) 32400 0.457 0.000 0.457 0.000 ops.py:4323(init) 32400 1.624 0.000 3.780 0.000 ops.py:4358(enter) 32400 1.115 0.000 3.377 0.000 ops.py:4368(exit) 32400 0.725 0.000 1.193 0.000 ops.py:4386(_push_control_dependencies_controller) 32400 1.010 0.000 1.749 0.000 ops.py:4389(_pop_control_dependencies_controller) 32400 0.596 0.000 1.053 0.000 ops.py:4443(control_dependencies) 64800 0.514 0.000 0.514 0.000 ops.py:4782(_device_function_stack) 64800 0.493 0.000 0.493 0.000 ops.py:4820(_device_function_stack) 129600 0.917 0.000 0.917 0.000 ops.py:4858(_control_dependencies_stack) 64800 0.514 0.000 0.514 0.000 ops.py:4870(_control_dependencies_stack) 32400 1.043 0.000 5.081 0.000 ops.py:4975(control_dependencies) 259200 3.878 0.000 5.612 0.000 ops.py:5015(get_default) 194400 3.084 0.000 4.587 0.000 ops.py:5032(get_controller) 32400 0.544 0.000 1.238 0.000 ops.py:5102(get_default_session) 226800 4.223 0.000 9.388 0.000 ops.py:5202(get_default) 32400 0.248 0.000 0.248 0.000 ops.py:5209(_GetGlobalDefaultGraph) 194400 12.833 0.000 45.151 0.000 ops.py:5222(get_controller) 64800 7.333 0.000 80.491 0.001 ops.py:5240(init_scope) 226800 3.500 0.000 12.888 0.000 ops.py:5545(get_default_graph) 32400 1.125 0.000 4.817 0.000 ops.py:5603(_get_graph_from_inputs) 32400 0.841 0.000 1.305 0.000 ops.py:5968(init) 32400 1.842 0.000 21.062 0.001 ops.py:5982(enter) 32400 0.957 0.000 8.383 0.000 ops.py:6033(exit) 1 0.000 0.000 662.242 662.242 profile:0(g2 = api.Game( copy.copy(algorithm_tuple), copy.copy(tokens) )) 0 0.000 0.000 profile:0(profiler) 32400 0.488 0.000 1.062 0.000 pywrap_tensorflow_internal.py:2124(Flatten) 1 0.000 0.000 0.000 0.000 random.py:175(randrange) 1 0.000 0.000 0.000 0.000 random.py:238(randint) 3780 0.082 0.000 0.131 0.000 random.py:273(choice) 26430 11.167 0.000 19.962 0.001 random.py:277(shuffle) 32400 0.979 0.000 27.716 0.001 sequential.py:110(layers) 32400 2.473 0.000 21.391 0.001 session.py:1429(call) 32400 0.234 0.000 0.234 0.000 session.py:746(graph) 64800 0.988 0.000 1.576 0.000 tape.py:68(stop_recording) 162000 4.980 0.000 11.698 0.000 tensor_util.py:941(is_tensor) 32400 0.775 0.000 1.422 0.000 tf_decorator.py:104(unwrap) 32400 0.845 0.000 2.986 0.000 tf_inspect.py:328(isgenerator) 32400 2.252 0.000 292.620 0.009 training.py:1778(predict) 32400 0.527 0.000 0.780 0.000 training.py:731(_make_predict_function) 32400 2.433 0.000 60.882 0.002 training.py:861(_standardize_user_data) 32400 2.501 0.000 46.948 0.001 training.py:995(_standardize_weights) 32400 9.064 0.000 224.220 0.007 training_arrays.py:246(predict_loop) 32400 2.892 0.000 5.738 0.000 training_utils.py:1018(init) 32400 0.224 0.000 0.224 0.000 training_utils.py:1100(as_list) 32400 0.886 0.000 10.080 0.000 training_utils.py:173(check_num_samples) 32400 0.542 0.000 0.762 0.000 training_utils.py:215(standardize_single_array) 32400 4.291 0.000 10.378 0.000 training_utils.py:226(standardize_input_data) 64800 1.464 0.000 12.648 0.000 training_utils.py:739(has_symbolic_tensors) 64800 1.786 0.000 8.696 0.000 training_utils.py:745(has_tensors) 64800 0.764 0.000 3.142 0.000 training_utils.py:747() 64800 2.783 0.000 17.017 0.000 training_utils.py:860(check_steps_argument) 32400 0.274 0.000 0.274 0.000 weakref.py:314(getitem) 32400 0.494 0.000 0.726 0.000 weakref.py:342(get) 32400 0.350 0.000 0.350 0.000 weakref.py:352(contains) 32400 0.540 0.000 0.815 0.000 weakref.py:432(setdefault)

[TheodoreEhrenborg]

Here's the profile of total_winner:

profile.run( "g22 = api.Game( copy.copy(algorithm_tuple2), copy.copy(tokens) )") 28264916 function calls (28261306 primitive calls) in 655.511 seconds

Ordered by: standard name

ncalls tottime percall cumtime percall filename:lineno(function) 19894748 215.765 0.000 215.765 0.000 :0(append) 140 0.004 0.000 0.004 0.000 :0(asctime) 140 0.035 0.000 0.035 0.000 :0(close) 54680 0.474 0.000 0.474 0.000 :0(count) 278332 2.069 0.000 2.069 0.000 :0(get) 54692 0.438 0.000 0.438 0.000 :0(getattr) 5650 0.043 0.000 0.043 0.000 :0(id) 52800 0.625 0.000 0.625 0.000 :0(index) 54752 1.024 0.000 4.533 0.000 :0(isinstance) 1597388 17.108 0.000 17.108 0.000 :0(len) 107480 1.214 0.000 1.214 0.000 :0(max) 52800 0.631 0.000 0.631 0.000 :0(min) 140 0.013 0.000 0.013 0.000 :0(open) 61 0.001 0.000 0.001 0.000 :0(pop) 2497441 17.494 0.000 17.494 0.000 :0(random) 2813104 33.329 0.000 33.329 0.000 :0(range) 5 0.000 0.000 0.000 0.000 :0(repr) 1 0.000 0.000 0.000 0.000 :0(setprofile) 3 0.000 0.000 0.000 0.000 :0(time) 140 0.003 0.000 0.003 0.000 :0(write) 1 0.000 0.000 655.511 655.511 :1() 109384 1.019 0.000 1.019 0.000 _weakrefset.py:70(contains) 54692 2.052 0.000 3.509 0.000 abc.py:128(instancecheck) 54680 26.856 0.000 62.762 0.001 algorithms.py:697(aux_abridged_game) 404800 4.291 0.000 4.291 0.000 algorithms.py:779(aux_list_total) 60 0.037 0.001 655.058 10.918 algorithms.py:794(total_winner) 60 301.901 5.032 655.021 10.917 algorithms.py:796(aux_stochastic) 1 0.024 0.024 655.511 655.511 api.py:449(init) 185 0.001 0.000 0.001 0.000 api.py:569(get_name) 60 0.000 0.000 0.000 0.000 api.py:571(get_function) 5 0.000 0.000 0.000 0.000 api.py:573(get_tuple) 5 0.000 0.000 0.000 0.000 api.py:575(repr) 65 0.001 0.000 0.001 0.000 api.py:577(str) 54692 1.757 0.000 13.266 0.000 collections.py:446(init) 54692 4.146 0.000 11.143 0.000 collections.py:516(update) 1 0.000 0.000 0.000 0.000 copy.py:101(_copy_immutable) 61 0.001 0.000 0.001 0.000 copy.py:113(_copy_with_constructor) 3130/120 0.128 0.000 0.353 0.003 copy.py:145(deepcopy) 960 0.006 0.000 0.006 0.000 copy.py:198(_deepcopy_atomic) 720/120 0.064 0.000 0.339 0.003 copy.py:226(_deepcopy_list) 1680 0.036 0.000 0.058 0.000 copy.py:267(_keep_alive) 62 0.002 0.000 0.003 0.000 copy.py:66(copy) 1 0.000 0.000 655.511 655.511 profile:0(g22 = api.Game( copy.copy(algorithm_tuple2), copy.copy(tokens) )) 0 0.000 0.000 profile:0(profiler) 1 0.000 0.000 0.000 0.000 random.py:175(randrange) 1 0.000 0.000 0.000 0.000 random.py:238(randint) 7560 0.170 0.000 0.270 0.000 random.py:273(choice) 52860 22.747 0.000 40.599 0.001 random.py:277(shuffle)

[TheodoreEhrenborg]

See https://hplgit.github.io/teamods/MC_cython/main_MC_cython.html https://cython.readthedocs.io/en/latest/src/tutorial/numpy.html

[TheodoreEhrenborg]

It doesn't seem to be working.

[TheodoreEhrenborg]

It's working slowly.

[TheodoreEhrenborg]

See https://docs.scipy.org/doc/numpy/reference/generated/numpy.bincount.html for a Collections alternative

[TheodoreEhrenborg]

This is important and useful: https://cython.readthedocs.io/en/latest/src/userguide/numpy_tutorial.html

[TheodoreEhrenborg]

The problem is that the computer has to look in the dictionary at least 100000 times. Using Cython does not speed this up, and replacing the Python dictionary with a custom array makes the program even slower. This problem also affects the stochastic programs besides neural_nash.