Avaliação de núcleos de CPU para otimização - página 7

 

Assim, mais uma vez, recompilou os três EAs.

Intel Core i7-8700, 3.2GHz, 16Gb RAM, 6 núcleos.

Hiperpreensão habilitada.

Seis agentes,Tree_Brut_TestPL:

2019.08.13 16:19:02.912 Core 03 pass 2 returned result 1001000.00 in 0:00:33.050
2019.08.13 16:19:02.943 Core 01 pass 0 returned result 1001000.00 in 0:00:33.127
2019.08.13 16:19:03.032 Core 02 pass 5 returned result 1001000.00 in 0:00:33.166
2019.08.13 16:19:03.238 Core 06 pass 4 returned result 1001000.00 in 0:00:33.372
2019.08.13 16:19:03.565 Core 04 pass 3 returned result 1001000.00 in 0:00:33.703
2019.08.13 16:19:03.784 Core 11 pass 1 returned result 1001000.00 in 0:00:33.947

Doze agentes,Tree_Brut_TestPL:

2019.08.13 16:24:28.900 Core 12 pass 11 returned result 1001000.00 in 0:01:43.168
2019.08.13 16:24:29.595 Core 05 pass 10 returned result 1001000.00 in 0:01:43.898
2019.08.13 16:24:29.897 Core 03 pass 2 returned result 1001000.00 in 0:01:44.675
2019.08.13 16:24:30.425 Core 10 pass 8 returned result 1001000.00 in 0:01:44.806
2019.08.13 16:24:30.488 Core 02 pass 1 returned result 1001000.00 in 0:01:45.264
2019.08.13 16:24:30.635 Core 09 pass 5 returned result 1001000.00 in 0:01:45.405
2019.08.13 16:24:30.720 Core 08 pass 7 returned result 1001000.00 in 0:01:45.122
2019.08.13 16:24:30.721 Core 06 pass 4 returned result 1001000.00 in 0:01:45.497
2019.08.13 16:24:31.081 Core 07 pass 9 returned result 1001000.00 in 0:01:45.422
2019.08.13 16:24:31.882 Core 01 pass 0 returned result 1001000.00 in 0:01:46.660
2019.08.13 16:24:31.992 Core 11 pass 6 returned result 1001000.00 in 0:01:46.769
2019.08.13 16:24:32.075 Core 04 pass 3 returned result 1001000.00 in 0:01:46.852
2019.08.13 16:24:32.075 Tester  optimization finished, total passes 12
2019.08.13 16:24:32.086 Statistics      optimization done in 1 minutes 47 seconds
2019.08.13 16:24:32.086 Statistics      shortest pass 0:01:43.168, longest pass 0:01:46.852, average pass 0:01:45.294

Seis agentes,Tree_Brut_TestPL_F:

2019.08.13 16:27:35.754 Core 01 pass 0 returned result 1001000.00 in 0:01:16.198
2019.08.13 16:27:35.969 Core 06 pass 5 returned result 1001000.00 in 0:01:16.423
2019.08.13 16:27:36.112 Core 04 pass 3 returned result 1001000.00 in 0:01:16.577
2019.08.13 16:27:36.246 Core 03 pass 2 returned result 1001000.00 in 0:01:16.764
2019.08.13 16:27:36.374 Core 02 pass 1 returned result 1001000.00 in 0:01:16.905
2019.08.13 16:27:37.184 Core 05 pass 4 returned result 1001000.00 in 0:01:17.664
2019.08.13 16:27:37.184 Tester  optimization finished, total passes 6
2019.08.13 16:27:37.194 Statistics      optimization done in 1 minutes 19 seconds
2019.08.13 16:27:37.194 Statistics      shortest pass 0:01:16.198, longest pass 0:01:17.664, average pass 0:01:16.755

Doze agentes,Árvore_Brut_TestPL_F:

2019.08.13 16:36:28.083 Core 11 pass 7 returned result 1001000.00 in 0:03:18.279
2019.08.13 16:36:28.473 Core 05 pass 5 returned result 1001000.00 in 0:03:18.672
2019.08.13 16:36:28.525 Core 01 pass 0 returned result 1001000.00 in 0:03:18.724
2019.08.13 16:36:29.260 Core 04 pass 2 returned result 1001000.00 in 0:03:19.459
2019.08.13 16:36:30.106 Core 10 pass 10 returned result 1001000.00 in 0:03:20.305
2019.08.13 16:36:30.865 Core 08 pass 6 returned result 1001000.00 in 0:03:21.064
2019.08.13 16:36:31.414 Core 06 pass 8 returned result 1001000.00 in 0:03:21.613
2019.08.13 16:36:31.556 Core 12 pass 11 returned result 1001000.00 in 0:03:21.754
2019.08.13 16:36:31.894 Core 03 pass 4 returned result 1001000.00 in 0:03:22.094
2019.08.13 16:36:33.473 Core 02 pass 1 returned result 1001000.00 in 0:03:23.673
2019.08.13 16:36:35.412 Core 09 pass 9 returned result 1001000.00 in 0:03:25.612
2019.08.13 16:36:35.493 Core 07 pass 3 returned result 1001000.00 in 0:03:25.693
2019.08.13 16:36:35.493 Tester  optimization finished, total passes 12
2019.08.13 16:36:35.503 Statistics      optimization done in 3 minutes 26 seconds
2019.08.13 16:36:35.503 Statistics      shortest pass 0:03:18.279, longest pass 0:03:25.693, average pass 0:03:21.411

Seis agentes,Árvore_Brut_TestPL_F_F_Fast:

2019.08.13 16:39:28.522 Core 02 pass 1 returned result 1001000.00 in 0:00:16.755
2019.08.13 16:39:28.549 Core 03 pass 2 returned result 1001000.00 in 0:00:16.781
2019.08.13 16:39:29.010 Core 06 pass 5 returned result 1001000.00 in 0:00:17.247
2019.08.13 16:39:29.085 Core 01 pass 0 returned result 1001000.00 in 0:00:17.375
2019.08.13 16:39:29.607 Core 04 pass 3 returned result 1001000.00 in 0:00:17.821
2019.08.13 16:39:29.716 Core 05 pass 4 returned result 1001000.00 in 0:00:17.996
2019.08.13 16:39:29.716 Tester  optimization finished, total passes 6
2019.08.13 16:39:29.726 Statistics      optimization done in 0 minutes 18 seconds
2019.08.13 16:39:29.726 Statistics      shortest pass 0:00:16.755, longest pass 0:00:17.996, average pass 0:00:17.329

Doze agentes, Tree_Brut_TestPL_F_F_Fast:

2019.08.13 16:41:11.638 Core 04 pass 3 returned result 1001000.00 in 0:00:29.382
2019.08.13 16:41:11.827 Core 06 pass 5 returned result 1001000.00 in 0:00:29.573
2019.08.13 16:41:12.313 Core 02 pass 1 returned result 1001000.00 in 0:00:30.060
2019.08.13 16:41:12.826 Core 07 pass 9 returned result 1001000.00 in 0:00:30.030
2019.08.13 16:41:12.907 Core 08 pass 6 returned result 1001000.00 in 0:00:30.115
2019.08.13 16:41:13.695 Core 01 pass 0 returned result 1001000.00 in 0:00:31.442
2019.08.13 16:41:13.927 Core 05 pass 4 returned result 1001000.00 in 0:00:31.673
2019.08.13 16:41:13.964 Core 09 pass 7 returned result 1001000.00 in 0:00:31.219
2019.08.13 16:41:15.006 Core 03 pass 2 returned result 1001000.00 in 0:00:32.752
2019.08.13 16:41:15.381 Core 10 pass 10 returned result 1001000.00 in 0:00:32.558
2019.08.13 16:41:16.244 Core 11 pass 8 returned result 1001000.00 in 0:00:33.420
2019.08.13 16:41:16.365 Core 12 pass 11 returned result 1001000.00 in 0:00:33.536
2019.08.13 16:41:16.365 Tester  optimization finished, total passes 12
2019.08.13 16:41:16.376 Statistics      optimization done in 0 minutes 34 seconds
2019.08.13 16:41:16.376 Statistics      shortest pass 0:00:29.382, longest pass 0:00:33.536, average pass 0:00:31.313

Desligando o hypertrading.

Seis agentes, Tree_Brut_TestPL:

2019.08.13 16:46:06.369 Core 5  pass 2 returned result 1001000.00 in 0:00:37.248
2019.08.13 16:46:06.958 Core 1  pass 0 returned result 1001000.00 in 0:00:37.861
2019.08.13 16:46:07.391 Core 6  pass 1 returned result 1001000.00 in 0:00:38.282
2019.08.13 16:46:07.722 Core 4  pass 4 returned result 1001000.00 in 0:00:38.611
2019.08.13 16:46:08.158 Core 3  pass 3 returned result 1001000.00 in 0:00:39.045
2019.08.13 16:46:08.305 Core 2  pass 5 returned result 1001000.00 in 0:00:39.164
2019.08.13 16:46:08.305 Tester  optimization finished, total passes 6
2019.08.13 16:46:08.316 Statistics      optimization done in 0 minutes 40 seconds
2019.08.13 16:46:08.316 Statistics      shortest pass 0:00:37.248, longest pass 0:00:39.164, average pass 0:00:38.368

Seis agentes, Tree_Brut_TestPL_F:

2019.08.13 16:53:21.112 Core 6  pass 3 returned result 1001000.00 in 0:01:10.789
2019.08.13 16:53:21.517 Core 1  pass 1 returned result 1001000.00 in 0:01:11.240
2019.08.13 16:53:21.660 Core 2  pass 0 returned result 1001000.00 in 0:01:11.523
2019.08.13 16:53:21.993 Core 4  pass 5 returned result 1001000.00 in 0:01:11.648
2019.08.13 16:53:22.054 Core 5  pass 4 returned result 1001000.00 in 0:01:11.713
2019.08.13 16:53:22.951 Core 3  pass 2 returned result 1001000.00 in 0:01:12.710
2019.08.13 16:53:22.951 Tester  optimization finished, total passes 6
2019.08.13 16:53:22.961 Statistics      optimization done in 1 minutes 13 seconds
2019.08.13 16:53:22.961 Statistics      shortest pass 0:01:10.789, longest pass 0:01:12.710, average pass 0:01:11.603

Seis agentes, Tree_Brut_TestPL_F_F_Fast:

2019.08.13 16:54:51.756 Core 3  pass 0 returned result 1001000.00 in 0:00:14.678
2019.08.13 16:54:51.810 Core 4  pass 4 returned result 1001000.00 in 0:00:14.727
2019.08.13 16:54:51.842 Core 2  pass 3 returned result 1001000.00 in 0:00:14.765
2019.08.13 16:54:51.873 Core 5  pass 1 returned result 1001000.00 in 0:00:14.881
2019.08.13 16:54:51.939 Core 6  pass 5 returned result 1001000.00 in 0:00:14.856
2019.08.13 16:54:51.961 Core 1  pass 2 returned result 1001000.00 in 0:00:14.915
2019.08.13 16:54:51.961 Tester  optimization finished, total passes 6
2019.08.13 16:54:51.972 Statistics      optimization done in 0 minutes 15 seconds
2019.08.13 16:54:51.972 Statistics      shortest pass 0:00:14.678, longest pass 0:00:14.915, average pass 0:00:14.803
 
Aleksey Vyazmikin:

Obrigado, mas acrescente Tree_Brut_TestPL_F e Tree_Brut_TestPL resultados para avaliação !

compila, é claro))) Só então eu vi a quantidade de código!

Árvore_Brut_TestPL

4 agentes 8 passes

Q       0       20:33:03.992    Core 3  pass 2 returned result 1001000.00 in 0:01:28.086
DF      0       20:33:06.162    Core 2  pass 0 returned result 1001000.00 in 0:01:30.305
DL      0       20:33:06.954    Core 1  pass 4 returned result 1001000.00 in 0:01:31.043
KE      0       20:33:06.998    Core 4  pass 6 returned result 1001000.00 in 0:01:31.025
EK      0       20:34:33.114    Core 3  pass 3 returned result 1001000.00 in 0:01:29.133
DP      0       20:34:36.197    Core 2  pass 1 returned result 1001000.00 in 0:01:30.045
GI      0       20:34:37.233    Core 4  pass 7 returned result 1001000.00 in 0:01:30.244
RO      0       20:34:37.410    Core 1  pass 5 returned result 1001000.00 in 0:01:30.465
KK      0       20:34:37.410    Tester  optimization finished, total passes 8
GO      0       20:34:37.420    Statistics      optimization done in 3 minutes 02 seconds
PF      0       20:34:37.420    Statistics      shortest pass 0:01:28.086, longest pass 0:01:31.043, average pass 0:01:30.043
CF      0       20:34:37.420    Statistics      8000 frames (3.14 Mb total, 412 bytes per frame) received
RE      0       20:34:37.421    Statistics      local 8 tasks (100%), remote 0 tasks (0%), cloud 0 tasks (0%)

8 agentes, 8 passes

EE      0       20:40:18.642    Core 1  pass 0 returned result 1001000.00 in 0:02:08.677
JJ      0       20:40:18.846    Core 8  pass 7 returned result 1001000.00 in 0:02:08.405
GP      0       20:40:19.243    Core 6  pass 5 returned result 1001000.00 in 0:02:08.930
DI      0       20:40:19.870    Core 2  pass 1 returned result 1001000.00 in 0:02:09.912
OO      0       20:40:20.049    Core 4  pass 3 returned result 1001000.00 in 0:02:10.090
MD      0       20:40:20.307    Core 3  pass 2 returned result 1001000.00 in 0:02:10.350
FM      0       20:40:21.323    Core 7  pass 6 returned result 1001000.00 in 0:02:10.983
PS      0       20:40:21.375    Core 5  pass 4 returned result 1001000.00 in 0:02:11.110
KG      0       20:40:21.375    Tester  optimization finished, total passes 8
NS      0       20:40:21.386    Statistics      optimization done in 2 minutes 12 seconds
NJ      0       20:40:21.386    Statistics      shortest pass 0:02:08.405, longest pass 0:02:11.110, average pass 0:02:09.807
RJ      0       20:40:21.386    Statistics      8000 frames (3.14 Mb total, 412 bytes per frame) received
DQ      0       20:40:21.386    Statistics      local 8 tasks (100%), remote 0 tasks (0%), cloud 0 tasks (0%)

Árvore_Brut_TestPL_F

4 agentes 8 passes

EI      0       20:47:44.732    Core 1  pass 0 returned result 1000000.00 in 0:02:43.903
MN      0       20:47:46.827    Core 4  pass 6 returned result 1000000.00 in 0:02:45.985
DG      0       20:47:49.535    Core 2  pass 2 returned result 1000000.00 in 0:02:48.698
GM      0       20:47:52.933    Core 3  pass 4 returned result 1000000.00 in 0:02:52.631
LR      0       20:50:13.736    Core 1  pass 1 returned result 1000000.00 in 0:02:29.026
EH      0       20:50:19.324    Core 4  pass 7 returned result 1000000.00 in 0:02:32.515
PQ      0       20:50:25.653    Core 2  pass 3 returned result 1000000.00 in 0:02:36.136
CF      0       20:50:32.200    Core 3  pass 5 returned result 1000000.00 in 0:02:39.285
ES      0       20:50:32.200    Tester  optimization finished, total passes 8
OG      0       20:50:32.211    Statistics      optimization done in 5 minutes 36 seconds
NN      0       20:50:32.211    Statistics      shortest pass 0:02:29.026, longest pass 0:02:52.631, average pass 0:02:41.022
LO      0       20:50:32.211    Statistics      8000 frames (3.14 Mb total, 412 bytes per frame) received
JJ      0       20:50:32.211    Statistics      local 8 tasks (100%), remote 0 tasks (0%), cloud 0 tasks (0%)

8 agentes 8 passes

KG      0       20:56:41.505    Core 3  pass 2 returned result 1000000.00 in 0:03:42.298
JL      0       20:56:42.220    Core 2  pass 1 returned result 1000000.00 in 0:03:43.017
IE      0       20:56:42.573    Core 1  pass 0 returned result 1000000.00 in 0:03:43.367
DK      0       20:56:43.809    Core 4  pass 3 returned result 1000000.00 in 0:03:44.609
DP      0       20:56:44.715    Core 7  pass 6 returned result 1000000.00 in 0:03:44.424
GF      0       20:56:45.115    Core 5  pass 4 returned result 1000000.00 in 0:03:44.974
IO      0       20:56:45.763    Core 6  pass 5 returned result 1000000.00 in 0:03:45.587
KD      0       20:56:46.010    Core 8  pass 7 returned result 1000000.00 in 0:03:45.727
GE      0       20:56:46.010    Tester  optimization finished, total passes 8
IQ      0       20:56:46.020    Statistics      optimization done in 3 minutes 48 seconds
JH      0       20:56:46.020    Statistics      shortest pass 0:03:42.298, longest pass 0:03:45.727, average pass 0:03:44.250
OM      0       20:56:46.020    Statistics      8000 frames (3.14 Mb total, 412 bytes per frame) received
FL      0       20:56:46.021    Statistics      local 8 tasks (100%), remote 0 tasks (0%), cloud 0 tasks (0%)

4 e 8 agentes 27-32% de diferença de velocidade. Eu acho que o problema está no cache de memória, FX é notório por seu gargalo de cache. Infelizmente não posso fazer overclock para verificá-lo, minha pedra não tem absolutamente nenhuma memória e o overclock da ponte norte.

 
mas se você comparar as versões f e f-f, a diferença é colossal! aqui elas são as maravilhas de 6,79 vezes otimização! Ou com qual versão deve ser rapidamente comparada?
 
Georgiy Merts:

Portanto, recompilou os três EAs mais uma vez.

Os números caíram um pouco - por enquanto os levou para a mesa. O computador permite o autoboosting durante a otimização?

 
Maxim Romanov:

4 e 8 agentes têm uma diferença de velocidade de 27-32 por cento. Eu acho que o problema aqui é o cache de memória, FX é famoso por seu gargalo de cache. Infelizmente não posso fazer overclock para verificá-lo, minha pedra não tem absolutamente nenhuma memória e o overclock da ponte norte.

Eu o escrevo para 4 núcleos de FPU.

Maxim Romanov:
mas se você comparar as versões f e f-ffast, a diferença é colossal! aqui elas são maravilhas da otimização 6,79! Ou com qual versão você deve comparar f-f-fast?

Concordo, é fenomenal. Você compara corretamente as versões. Curiosamente, a lacuna da i7-8700 diminuiu em termos percentuais para os processadores mais antigos.

 

Dados atuais - nenhum resultado de teste em azul.

A E5-2670 2pc está agora claramente à frente da i7-8700, o que faz sentido.

 
Aleksey Vyazmikin:

Os números caíram ligeiramente - por enquanto, os levou para uma mesa. O computador permite o aumento automático durante a otimização?

Muito provavelmente, sim. Há muitos parâmetros de afinação na BIOS, mas eu não entrei neles, apenas defini o perfil de desempenho.

 
Georgiy Merts:

Muito provavelmente, sim. Há muitos parâmetros de ajuste fino na BIOS, mas eu não entrei neles, eu apenas defini o perfil de desempenho.

Então, talvez quando o hypertrading for desligado, o impulsionador será ativado mais rapidamente ou em uma freqüência maior. Seria bom fixar a freqüência para testes objetivos.

 

Adicionei os resultados do processador E5-2470 - 2 pc, embora a memória lá seja DDR3-800, o que provavelmente subestima ligeiramente o potencial, mas não acho que isso seja significativo - 5% a 10%.

Decidi analisar os preços dos componentes usados para avaliar a viabilidade da montagem - os preços são retirados da Avito e da Aliexpress.

O cálculo foi feito da seguinte forma - o custo de montagem (coluna "Total") foi dividido pelo número de passes por hora. A filtragem é realizada para a última mas uma coluna, já que o Expert Advisor é o mais pesado e nem todos eles, infelizmente, deram os dados para a última EA.


Parece que a i7-8700 é uma opção realmente boa para otimização, mesmo levando em conta o preço de mercado usado.

A memória para processadores AMD no soquete AM3 e AM3+ foi tomada como especial, somente para AMD (eu mesmo a uso) - freqüência 1600.

Frequência de memória DDR4 2600 - maior freqüência não é comensurável com ganho de desempenho e, a julgar por testes de outras fontes.

Para um processador um pouco potente pode economizar no refrigerador - 500 rublos.

O fornecimento de energia elétrica levou 500 watts - isto é realmente suficiente, se não falso ou já morto, mas se possível melhor para obter um novo, mas não de ponta, mas o preço do bu é fixado em uma opção muito sólida.

A placa gráfica em todas as variantes está integrada no processador ou mãe, exceto Ryzen, mas o plugue em 250-300 rublos decidiu não levar em conta como uma exceção.

 

Parece que o assunto está apodrecendo, vou largar o arquivo que usei para contar e desenhar as tabelas em excel.

Arquivos anexados:
CPU_Test_FX.zip  25 kb