-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbenchmarks_other_libraries.jl
455 lines (382 loc) · 17.9 KB
/
benchmarks_other_libraries.jl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
# Benchmark with other Game Theory libraries: GameTheory.jl, Nashpy, pygambit
using Pkg
cd(@__DIR__)
Pkg.activate(".")
using BenchmarkTools
using StrategicGames
using GameTheory
using CSV, DataFrames
using PyCall
#ENV["JULIA_PYTHONCALL_EXE"] = "/home/lobianco/.pyenv/shims/python3"
#import PythonCall
#using Nash.jl # https://github.com/KrainskiL/Nash.jl
#import StrategicGames: nash_se2
const nash = PyCall.pyimport("nashpy") # Equiv. of Python `import nashpy as nash`
const pygambit = PyCall.pyimport("pygambit")
# ------------------------------------------------------------------------------
# Preliminary set up....
bms = DataFrame(benchmark_name = String[], library = String[], method = String[], time = Float64[], memory = Union{Float64,Missing}[], alloc = Union{Int64,Missing}[], neqs = Int64[], notes=String[])
macro binfo(bexpr)
runexpr = quote
b = @benchmark $bexpr
return (median(b).time, median(b).memory, median(b).allocs)
end
return eval(runexpr)
end
py"""
import numpy as np
import pygambit
def setGambitGame2pInt(P1,P2):
p1 = np.array(P1, dtype=pygambit.Rational)
p2 = np.array(P2, dtype=pygambit.Rational)
return pygambit.Game.from_arrays(p1,p2)
def setGambitGame3pInt(P1,P2,P3):
p1 = np.array(P1, dtype=pygambit.Rational)
p2 = np.array(P2, dtype=pygambit.Rational)
p3 = np.array(P3, dtype=pygambit.Rational)
return pygambit.Game.from_arrays(p1,p2,p3)
def setGambitGame2pDec(P1,P2):
p1 = np.array(P1, dtype=pygambit.Decimal)
p2 = np.array(P2, dtype=pygambit.Decimal)
print(p1)
return pygambit.Game.from_arrays(p1,p2)
"""
# For test
res = @binfo rand(2,2)
push!(bms,["Test","rand","rand(2,2)",res...,1,""])
CSV.write("bms.csv",bms)
# ------------------------------------------------------------------------------
# Test 1 - small 3x2 game
game_name = "small_3x2"
u = [(3,3) (3,2);
(2,2) (5,6);
(0,3) (6,1)]
payoff = expand_dimensions(u)
small_3x2_StrategicGames = payoff
small_3x2_GameTheory = NormalFormGame(payoff)
small_3x2_nash = nash.Game(payoff[:,:,1], payoff[:,:,2])
small_3x2_gambit = py"""setGambitGame2pInt"""(payoff[:,:,1],payoff[:,:,2])
eqs = nash_cp(small_3x2_StrategicGames,verbosity=NONE)
neqs = 1
res = @binfo nash_cp($small_3x2_StrategicGames,verbosity=NONE)
push!(bms,[game_name,"StrategicGames","nash_cp",res...,neqs,""])
eqs = nash_se(small_3x2_StrategicGames,max_samples=Inf, mt=true)
neqs = length(eqs)
res = @binfo nash_se($small_3x2_StrategicGames,max_samples=Inf)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,""])
eqs = hc_solve(small_3x2_GameTheory)
neqs = length(eqs)
res = @binfo hc_solve($small_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,""])
eqs = support_enumeration(small_3x2_GameTheory)
neqs = length(eqs)
res = @binfo support_enumeration($small_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","support_enumeration",res..., neqs,""])
eqs = lrsnash(small_3x2_GameTheory)
neqs = length(eqs)
res = @binfo lrsnash($small_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","lrsnash",res..., neqs,""])
eqs_gen = small_3x2_nash.vertex_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $small_3x2_nash.vertex_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","vertex_enumeration",res...,neqs,""])
eqs_gen = small_3x2_nash.lemke_howson_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $small_3x2_nash.lemke_howson_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","lemke_howson_enumeration",res...,neqs,"repeated results"])
eqs_gen = small_3x2_nash.support_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $small_3x2_nash.support_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","support_enumeration",res...,neqs,""])
eqs = pygambit.nash.lcp_solve(small_3x2_gambit)
neqs= length(eqs)
res = @binfo pygambit.nash.lcp_solve(small_3x2_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","lcp_solve",res...,neqs,""])
solver = pygambit.nash.ExternalEnumPolySolver()
eqs = solver.solve(small_3x2_gambit)
neqs= length(eqs)
res = @binfo solver.solve(small_3x2_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","ExternalEnumPolySolver",res...,neqs,""])
CSV.write("bms.csv",bms)
# ------------------------------------------------------------------------------
# Test 2 - mid rand 6x7 game
game_name = "rand_6x7"
#payoff = rand(1:1000,6,7,2)
p1 = [
253 646 641 395 258 153 375
713 17 145 582 338 258 145
174 265 282 588 80 996 478
346 517 963 829 976 735 334
492 106 199 986 278 658 407
506 288 439 345 549 869 986]
p2 = [
296 23 932 537 515 130 679
491 49 315 432 977 799 777
284 761 914 268 313 864 25
75 944 444 209 804 452 502
869 837 950 707 755 452 9
579 794 747 272 929 288 780
]
payoff = cat(p1,p2,dims=3)
rand_6x7_StrategicGames = payoff
rand_6x7_GameTheory = NormalFormGame(payoff)
rand_6x7_nash = nash.Game(payoff[:,:,1], payoff[:,:,2])
rand_6x7_gambit = py"""setGambitGame2pInt"""(payoff[:,:,1],payoff[:,:,2])
eqs = nash_cp(rand_6x7_StrategicGames)
neqs = 0
res = @binfo nash_cp($rand_6x7_StrategicGames,verbosity=NONE)
push!(bms,[game_name,"StrategicGames","nash_cp",res...,neqs,""])
eqs = nash_se(rand_6x7_StrategicGames,max_samples=Inf)
neqs = length(eqs)
res = @binfo nash_se($rand_6x7_StrategicGames,max_samples=Inf)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,""])
eqs = hc_solve(rand_6x7_GameTheory)
neqs = length(eqs)
res = @binfo hc_solve($rand_6x7_GameTheory)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,""])
eqs = support_enumeration(rand_6x7_GameTheory)
neqs = length(eqs)
res = @binfo support_enumeration($rand_6x7_GameTheory)
push!(bms,[game_name,"GameTheory","support_enumeration",res..., neqs,""])
eqs = lrsnash(rand_6x7_GameTheory)
neqs = length(eqs)
res = @binfo lrsnash($rand_6x7_GameTheory)
push!(bms,[game_name,"GameTheory","lrsnash",res..., neqs,""])
eqs_gen = rand_6x7_nash.vertex_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_6x7_nash.vertex_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","vertex_enumeration",res...,neqs,""])
eqs_gen = rand_6x7_nash.lemke_howson_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_6x7_nash.lemke_howson_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","lemke_howson_enumeration",res...,neqs,"repeated results"])
eqs_gen = rand_6x7_nash.support_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_6x7_nash.support_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","support_enumeration",res...,neqs,""])
CSV.write("bms.csv",bms)
eqs = pygambit.nash.lcp_solve(rand_6x7_gambit)
neqs= length(eqs)
res = @binfo pygambit.nash.lcp_solve(rand_6x7_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","lcp_solve",res...,neqs,""])
solver = pygambit.nash.ExternalEnumPolySolver()
eqs = solver.solve(rand_6x7_gambit) # takes a while !!!!
neqs= length(eqs)
#res = @binfo solver.solve(rand_6x7_gambit) # takes even more !
#res = (res[1], missing, missing)
res = (5.8840781273e11, missing, missing) # swap comments to run the test, takes ~ 2 hours on a good laptop
push!(bms,[game_name,"pygambit","ExternalEnumPolySolver",res...,neqs,""])
CSV.write("bms.csv",bms)
# ------------------------------------------------------------------------------
# Test 3 - rand decimal 6x5 game
game_name = "rand_dec_6x5"
# Note that I can't get Gambit working with decimal data, not even expressed as rational:
# This would both result in "TypeError: payoff argument should be a numeric type instance"
# g = pygambit.Game.from_arrays(np.array([[11/10,21/10,31/10],[41/10,51/10,61/10]], dtype = pygambit.Rational),np.array([[101/100,201/100,301/100],[401/100,501/100,601/600]], dtype = pygambit.Rational))
# g = pygambit.Game.from_arrays(np.array([[1.1,2.1,3.1],[4.1,5.1,6.1]], dtype = pygambit.Decimal),np.array([[10.1,20.1,30.1],[40.1,50.1,60.1]], dtype = pygambit.Decimal))
#payoff = rand(6,5,2)
p1 = [
0.843789 0.0777291 0.0265584 0.664203 0.716576
0.399443 0.356958 0.54122 0.817734 0.22748
0.346581 0.632591 0.712136 0.250977 0.797303
0.315566 0.861972 0.089729 0.801083 0.386999
0.895148 0.279619 0.83699 0.844866 0.871585
0.0835494 0.647902 0.0942865 0.405726 0.542989]
p2 = [
0.443203 0.810953 0.904374 0.270971 0.293723
0.527435 0.72811 0.4115 0.496543 0.453296
0.689995 0.612925 0.201368 0.578151 0.572826
0.0396185 0.722874 0.486394 0.100332 0.527595
0.0671089 0.318471 0.968878 0.0495921 0.803702
0.207152 0.848768 0.445008 0.293703 0.239207
]
payoff = cat(p1,p2,dims=3)
rand_dec_6x5_StrategicGames = payoff
rand_dec_6x5_GameTheory = NormalFormGame(payoff)
rand_dec_6x5_GameTheory_rat = NormalFormGame(Rational.(payoff))
rand_dec_6x5_nash = nash.Game(payoff[:,:,1], payoff[:,:,2])
eqs = nash_cp(rand_dec_6x5_StrategicGames)
neqs = 1
res = @binfo nash_cp($rand_dec_6x5_StrategicGames,verbosity=NONE)
push!(bms,[game_name,"StrategicGames","nash_cp",res...,neqs,""])
eqs = nash_se(rand_dec_6x5_StrategicGames,max_samples=Inf)
neqs = length(eqs)
res = @binfo nash_se($rand_dec_6x5_StrategicGames,max_samples=Inf)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,""])
eqs = hc_solve(rand_dec_6x5_GameTheory)
neqs = length(eqs)
res = @binfo hc_solve($rand_dec_6x5_GameTheory)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,""])
eqs = support_enumeration(rand_dec_6x5_GameTheory)
neqs = length(eqs)
res = @binfo support_enumeration($rand_dec_6x5_GameTheory)
push!(bms,[game_name,"GameTheory","support_enumeration",res..., neqs,""])
eqs = lrsnash(rand_dec_6x5_GameTheory_rat)
neqs = length(eqs)
res = @binfo lrsnash($rand_dec_6x5_GameTheory_rat)
push!(bms,[game_name,"GameTheory","lrsnash",res..., neqs,""])
eqs_gen = rand_dec_6x5_nash.vertex_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_dec_6x5_nash.vertex_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","vertex_enumeration",res...,neqs,""])
eqs_gen = rand_dec_6x5_nash.lemke_howson_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_dec_6x5_nash.lemke_howson_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","lemke_howson_enumeration",res...,neqs,"repeated results"])
eqs_gen = rand_dec_6x5_nash.support_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $rand_dec_6x5_nash.support_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","support_enumeration",res...,neqs,""])
CSV.write("bms.csv",bms)
# ------------------------------------------------------------------------------
# Test 4 - rand 4x4x2 game
game_name = "rand_4x4x2"
#payoff = rand(1:1000,4,4,2,3)
p1 = [656 207 580 548; 208 337 496 805; 640 422 152 691; 89 183 56 167;;; 44 789 965 925; 917 475 148 990; 78 563 732 25; 424 110 965 431]
p2 = [955 92 480 205; 436 135 802 301; 206 316 346 699; 912 470 317 787;;; 207 536 967 212; 990 335 573 61; 865 455 698 356; 200 141 978 848]
p3 = [364 963 199 81; 332 994 362 972; 421 780 852 150; 72 308 723 828;;; 737 785 653 960; 389 791 125 709; 421 483 10 404; 552 622 161 810]
payoff = cat(p1,p2,p3,dims=4)
rand_4x4x2_StrategicGames = payoff
rand_4x4x2_GameTheory = NormalFormGame(payoff)
rand_4x4x2_gambit = py"""setGambitGame3pInt"""(payoff[:,:,:,1],payoff[:,:,:,2],payoff[:,:,:,3])
eqs = nash_cp(rand_4x4x2_StrategicGames)
neqs = 0
res = @binfo nash_cp($rand_4x4x2_StrategicGames,verbosity=NONE)
push!(bms,[game_name,"StrategicGames","nash_cp",res...,neqs,""])
eqs = nash_se(rand_4x4x2_StrategicGames,max_samples=Inf)
neqs = length(eqs)
res = @binfo nash_se($rand_4x4x2_StrategicGames,max_samples=Inf)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,"1 eq repeated"])
eqs = hc_solve(rand_4x4x2_GameTheory)
neqs = length(eqs)
res = @binfo hc_solve($rand_4x4x2_GameTheory)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,"2 eq missing"])
CSV.write("bms.csv",bms)
solver = pygambit.nash.ExternalEnumPolySolver()
eqs = solver.solve(rand_4x4x2_gambit)
neqs= length(eqs)
res = @binfo solver.solve(rand_4x4x2_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","ExternalEnumPolySolver",res...,neqs,"1 eq missed"])
CSV.write("bms.csv",bms)
# s = [[0.5531496062992137, 0.0, 0.0, 0.4468503937007864], [0.0, 0.0, 1.0, 0.0], [-9.755303440999721e-15, 1.0000000000000098]]
# StrategicGames.is_nash(payoff,s)
# ------------------------------------------------------------------------------
# Test 5 - mid rand 6x7 game first solution only
game_name = "rand_6x7_1st_eq"
#payoff = rand(1:1000,6,7,2)
p1 = [
253 646 641 395 258 153 375
713 17 145 582 338 258 145
174 265 282 588 80 996 478
346 517 963 829 976 735 334
492 106 199 986 278 658 407
506 288 439 345 549 869 986]
p2 = [
296 23 932 537 515 130 679
491 49 315 432 977 799 777
284 761 914 268 313 864 25
75 944 444 209 804 452 502
869 837 950 707 755 452 9
579 794 747 272 929 288 780
]
payoff = cat(p1,p2,dims=3)
rand_6x7_StrategicGames = payoff
rand_6x7_GameTheory = NormalFormGame(payoff)
rand_6x7_nash = nash.Game(payoff[:,:,1], payoff[:,:,2])
rand_6x7_gambit = py"""setGambitGame2pInt"""(payoff[:,:,1],payoff[:,:,2])
eqs = nash_se(rand_6x7_StrategicGames,max_samples=1,mt=true)
neqs = length(eqs)
res = @binfo nash_se($rand_6x7_StrategicGames,max_samples=1,mt=true)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,""])
eqs = hc_solve(rand_6x7_GameTheory,ntofind=1)
neqs = length(eqs)
res = @binfo hc_solve($rand_6x7_GameTheory,ntofind=1)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,""])
eqs_gen = rand_6x7_nash.vertex_enumeration(); eqs = py"next($eqs_gen)"
neqs= 1
res = @binfo begin eqs = $rand_6x7_nash.vertex_enumeration(); py"next($eqs)" end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","vertex_enumeration",res...,neqs,""])
eqs_gen = rand_6x7_nash.lemke_howson_enumeration(); eqs = py"next($eqs_gen)"
neqs= 1
res = @binfo begin eqs = $rand_6x7_nash.lemke_howson_enumeration(); py"next($eqs)" end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","lemke_howson_enumeration",res...,neqs,""])
eqs_gen = rand_6x7_nash.support_enumeration();
# eqs = py"next($eqs_gen)"
neqs= 0
res = @binfo begin eqs = $rand_6x7_nash.support_enumeration(); end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","support_enumeration",res...,neqs,"no eq reported"])
CSV.write("bms.csv",bms)
# ------------------------------------------------------------------------------
# Test 6 - degenerated 3x2 game
game_name = "degenerated_3x2"
u = [(3,3) (3,3);
(2,2) (5,6);
(0,3) (6,1)]
payoff = expand_dimensions(u)
degenerated_3x2_StrategicGames = payoff
degenerated_3x2_GameTheory = NormalFormGame(payoff)
degenerated_3x2_nash = nash.Game(payoff[:,:,1], payoff[:,:,2])
degenerated_3x2_gambit = py"""setGambitGame2pInt"""(payoff[:,:,1],payoff[:,:,2])
eqs = nash_cp(degenerated_3x2_StrategicGames,verbosity=NONE)
neqs = 1
res = @binfo nash_cp($degenerated_3x2_StrategicGames,verbosity=NONE)
push!(bms,[game_name,"StrategicGames","nash_cp",res...,neqs,""])
eqs = nash_se(degenerated_3x2_StrategicGames,max_samples=Inf, mt=true)
neqs = length(eqs)
res = @binfo nash_se($degenerated_3x2_StrategicGames,max_samples=Inf)
push!(bms,[game_name,"StrategicGames","nash_se",res...,neqs,""])
eqs = hc_solve(degenerated_3x2_GameTheory)
neqs = length(eqs)
res = @binfo hc_solve($degenerated_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","hc_solve",res..., neqs,""])
eqs = support_enumeration(degenerated_3x2_GameTheory)
neqs = length(eqs)
res = @binfo support_enumeration($degenerated_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","support_enumeration",res..., neqs,""])
eqs = lrsnash(degenerated_3x2_GameTheory)
neqs = length(eqs)
res = @binfo lrsnash($degenerated_3x2_GameTheory)
push!(bms,[game_name,"GameTheory","lrsnash",res..., neqs,""])
eqs_gen = degenerated_3x2_nash.vertex_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $degenerated_3x2_nash.vertex_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","vertex_enumeration",res...,neqs,""])
eqs_gen = degenerated_3x2_nash.lemke_howson_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $degenerated_3x2_nash.lemke_howson_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","lemke_howson_enumeration",res...,neqs,"2 repeated results"])
eqs_gen = degenerated_3x2_nash.support_enumeration(); eqs = [eq for eq in eqs_gen]
neqs= length(eqs)
res = @binfo begin eqs = $degenerated_3x2_nash.support_enumeration(); [eq for eq in eqs] end
res = (res[1], missing, missing)
push!(bms,[game_name,"nashpy","support_enumeration",res...,neqs,""])
eqs = pygambit.nash.lcp_solve(degenerated_3x2_gambit)
neqs= length(eqs)
res = @binfo pygambit.nash.lcp_solve(degenerated_3x2_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","lcp_solve",res...,neqs,""])
solver = pygambit.nash.ExternalEnumPolySolver()
eqs = solver.solve(degenerated_3x2_gambit)
neqs= length(eqs)
res = @binfo solver.solve(degenerated_3x2_gambit)
res = (res[1], missing, missing)
push!(bms,[game_name,"pygambit","ExternalEnumPolySolver",res...,neqs,""])
CSV.write("bms.csv",bms)