KataGo线程/权重/显卡

权重: 40b
显卡: 580 Series
线程: 8:: EloDiff: +78    Visits/s: 200.65    nnEvals/s: 148.19    nnBatches/s: 37.07    avgBatchSize: 4.00(249.5secs)


权重: 40b
显卡: 2060 SUPER
线程: 12:: EloDiff: +134    Visits/s: 726.57    nnEvals/s: 511.66    nnBatches/s: 85.94    avgBatchSize: 5.95(69.0secs)
线程: 13:: EloDiff: +143    Visits/s: 744.08    nnEvals/s: 523.72    nnBatches/s: 81.19    avgBatchSize: 6.45(67.4secs)


权重: 40b
显卡: 1080 Ti
线程: 20:: EloDiff: +87    Visits/s: 423.08    nnEvals/s: 303.32    nnBatches/s: 30.42    avgBatchSize: 9.97(177.7secs)


权重: 40b
显卡: 2080
线程: 26:: EloDiff: +124    Visits/s: 1047.61    nnEvals/s: 726.87    nnBatches/s: 42.00    avgBatchSize: 17.31(71.8secs)


权重: 40b
显卡: 3060 Ti
线程: 36:: EloDiff: +206    Visits/s: 1390.41    nnEvals/s: 913.15    nnBatches/s: 51.70    avgBatchSize: 17.66(72.2secs)
线程: 32:: EloDiff: +147    Visits/s: 641.54    nnEvals/s: 458.08    nnBatches/s: 29.32    avgBatchSize: 15.63(78.4secs)
线程: 24:: EloDiff: +169    Visits/s: 1296.42    nnEvals/s: 923.84    nnBatches/s: 77.18    avgBatchSize: 11.97(77.3secs)


权重: 40b
显卡: 2080 Ti, 2070 SUPER, 2070 SUPER, 2080 Ti
线程: 69:: EloDiff: +430    Visits/s: 2940.53    nnEvals/s: 2220.42    nnBatches/s: 299.27    avgBatchSize: 7.42(17.2secs)
线程: 62:: EloDiff: +441    Visits/s: 2969.02    nnEvals/s: 2181.68    nnBatches/s: 326.18    avgBatchSize: 6.69(17.0secs)
线程: 68:: EloDiff: +372    Visits/s: 2961.77    nnEvals/s: 2225.45    nnBatches/s: 289.34    avgBatchSize: 7.69(17.1secs)
线程: 70:: EloDiff: +401    Visits/s: 3341.24    nnEvals/s: 2486.65    nnBatches/s: 311.78    avgBatchSize: 7.98(15.2secs)


权重: 40b
显卡: Tesla V100-PCIE-16GB, TITAN V
线程: 158:: EloDiff: +695    Visits/s: 4007.81    nnEvals/s: 2915.87    nnBatches/s: 83.02    avgBatchSize: 35.12(25.3secs)


权重: 40b
显卡: 3060 Laptop GPU
线程: 24:: EloDiff: +158    Visits/s: 952.70    nnEvals/s: 724.56    nnBatches/s: 61.42    avgBatchSize: 11.80(33.8secs)


权重: 40b
显卡: 2070 SUPER
线程: 24:: EloDiff: +181    Visits/s: 1066.93    nnEvals/s: 661.64    nnBatches/s: 55.83    avgBatchSize: 11.85(234.5secs)


权重: 40b
显卡: 2060
线程: 25:: EloDiff: +210    Visits/s: 612.62    nnEvals/s: 434.84    nnBatches/s: 24.75    avgBatchSize: 17.57(98.3secs)
线程: 16:: EloDiff: +198    Visits/s: 565.27    nnEvals/s: 406.51    nnBatches/s: 45.13    avgBatchSize: 9.01(88.7secs)
线程: 17:: EloDiff: +182    Visits/s: 584.34    nnEvals/s: 404.65    nnBatches/s: 44.87    avgBatchSize: 9.02(103.0secs)
线程: 11:: EloDiff: +119    Visits/s: 567.73    nnEvals/s: 390.14    nnBatches/s: 71.29    avgBatchSize: 5.47(105.9secs)
线程: 17:: EloDiff: +154    Visits/s: 569.82    nnEvals/s: 401.31    nnBatches/s: 43.61    avgBatchSize: 9.20(105.6secs)
线程: 18:: EloDiff: +176    Visits/s: 593.29    nnEvals/s: 408.22    nnBatches/s: 41.48    avgBatchSize: 9.84(84.6secs)
线程: 16:: EloDiff: +111    Visits/s: 588.37    nnEvals/s: 407.44    nnBatches/s: 49.57    avgBatchSize: 8.22(85.2secs)
线程: 11:: EloDiff: +96    Visits/s: 567.11    nnEvals/s: 393.49    nnBatches/s: 71.87    avgBatchSize: 5.48(88.3secs)
线程: 17:: EloDiff: +185    Visits/s: 573.25    nnEvals/s: 404.50    nnBatches/s: 44.85    avgBatchSize: 9.02(87.5secs)
线程: 15:: EloDiff: +98    Visits/s: 593.38    nnEvals/s: 405.53    nnBatches/s: 53.59    avgBatchSize: 7.57(84.5secs)
线程: 12:: EloDiff: +95    Visits/s: 571.46    nnEvals/s: 395.97    nnBatches/s: 66.27    avgBatchSize: 5.98(87.7secs)
线程: 18:: EloDiff: +179    Visits/s: 567.08    nnEvals/s: 398.32    nnBatches/s: 39.29    avgBatchSize: 10.14(88.5secs)
线程: 12:: EloDiff: +113    Visits/s: 564.62    nnEvals/s: 398.29    nnBatches/s: 66.63    avgBatchSize: 5.98(88.8secs)
线程: 16:: EloDiff: +185    Visits/s: 563.98    nnEvals/s: 396.05    nnBatches/s: 47.93    avgBatchSize: 8.26(88.9secs)
线程: 13:: EloDiff: +99    Visits/s: 566.51    nnEvals/s: 399.77    nnBatches/s: 61.14    avgBatchSize: 6.54(88.5secs)
线程: 13:: EloDiff: +95    Visits/s: 560.29    nnEvals/s: 397.90    nnBatches/s: 60.76    avgBatchSize: 6.55(89.5secs)
线程: 12:: EloDiff: +106    Visits/s: 550.19    nnEvals/s: 396.37    nnBatches/s: 65.70    avgBatchSize: 6.03(91.1secs)
线程: 16:: EloDiff: +216    Visits/s: 573.25    nnEvals/s: 397.09    nnBatches/s: 48.37    avgBatchSize: 8.21(87.5secs)
线程: 18:: EloDiff: +100    Visits/s: 581.90    nnEvals/s: 413.70    nnBatches/s: 40.84    avgBatchSize: 10.13(86.2secs)
线程: 18:: EloDiff: +176    Visits/s: 563.13    nnEvals/s: 404.44    nnBatches/s: 41.24    avgBatchSize: 9.81(89.1secs)
线程: 17:: EloDiff: +175    Visits/s: 562.50    nnEvals/s: 401.86    nnBatches/s: 44.87    avgBatchSize: 8.96(89.2secs)
线程: 15:: EloDiff: +195    Visits/s: 548.91    nnEvals/s: 399.82    nnBatches/s: 48.32    avgBatchSize: 8.27(91.3secs)
线程: 10:: EloDiff: +167    Visits/s: 597.75    nnEvals/s: 423.57    nnBatches/s: 84.81    avgBatchSize: 4.99(83.8secs)
线程: 10:: EloDiff: +153    Visits/s: 589.84    nnEvals/s: 429.21    nnBatches/s: 85.97    avgBatchSize: 4.99(84.9secs)
线程: 10:: EloDiff: +151    Visits/s: 582.54    nnEvals/s: 427.28    nnBatches/s: 85.58    avgBatchSize: 4.99(86.0secs)
线程: 16:: EloDiff: +214    Visits/s: 562.99    nnEvals/s: 400.99    nnBatches/s: 44.84    avgBatchSize: 8.94(89.1secs)
线程: 10:: EloDiff: +153    Visits/s: 587.36    nnEvals/s: 422.13    nnBatches/s: 84.31    avgBatchSize: 5.01(85.3secs)
线程: 10:: EloDiff: +133    Visits/s: 603.28    nnEvals/s: 432.82    nnBatches/s: 86.66    avgBatchSize: 4.99(83.0secs)


权重: 40b
显卡: 2080 Ti
线程: 26:: EloDiff: +240    Visits/s: 695.58    nnEvals/s: 495.21    nnBatches/s: 31.88    avgBatchSize: 15.53(287.9secs)


权重: 40b
显卡: 2080 Ti, 2080 Ti
线程: 54:: EloDiff: +337    Visits/s: 3042.62    nnEvals/s: 1873.08    nnBatches/s: 112.60    avgBatchSize: 16.64(65.9secs)
线程: 20:: EloDiff: +168 Visits/s: 1018.78    nnEvals/s: 637.74    nnBatches/s: 103.49    avgBatchSize: 6.16(196.5secs)
线程: 56:: EloDiff: +338    Visits/s: 3027.22    nnEvals/s: 1882.31    nnBatches/s: 110.34    avgBatchSize: 17.06(66.2secs)
线程: 45:: EloDiff: +321    Visits/s: 2869.87    nnEvals/s: 1814.13    nnBatches/s: 129.32    avgBatchSize: 14.03(69.8secs)


权重: 40b
显卡: 3080
线程: 36:: EloDiff: +197    Visits/s: 1942.66    nnEvals/s: 1271.76    nnBatches/s: 72.21    avgBatchSize: 17.61(51.7secs)
线程: 36:: EloDiff: +197    Visits/s: 1918.22    nnEvals/s: 1278.55    nnBatches/s: 72.43    avgBatchSize: 17.65(52.3secs)


权重: 40b
显卡: 2060, 2060
线程: 32:: EloDiff: +191    Visits/s: 1484.10    nnEvals/s: 1010.33    nnBatches/s: 102.23    avgBatchSize: 9.88(54.7secs)


权重: 40b
显卡: 2060, 2060, 2060
线程: 28:: EloDiff: +292    Visits/s: 1611.06    nnEvals/s: 1157.16    nnBatches/s: 240.71    avgBatchSize: 4.81(31.2secs)
线程: 48:: EloDiff: +355    Visits/s: 2091.69    nnEvals/s: 1339.69    nnBatches/s: 163.80    avgBatchSize: 8.18(71.9secs)
线程: 96:: EloDiff: +323    Visits/s: 2238.39    nnEvals/s: 1489.29    nnBatches/s: 89.09    avgBatchSize: 16.72(67.4secs)


权重: 40b
显卡: 750 Ti
线程: 8:: EloDiff: +63    Visits/s: 71.30    nnEvals/s: 52.32    nnBatches/s: 13.14    avgBatchSize: 3.98(449.8secs)

权重: kata1-b40c256-s790.gz
显卡: 2080
线程: 30:: EloDiff: +215    Visits/s: 1031.16    nnEvals/s: 712.48    nnBatches/s: 37.70    avgBatchSize: 18.90(73.0secs)


权重: kata1-b40c256-s7850687232-d1904191439.bin.gz
显卡: 2080, 2060 SUPER
线程: 38:: EloDiff: +288    Visits/s: 1940.54    nnEvals/s: 1239.80    nnBatches/s: 109.22    avgBatchSize: 11.35(54.8secs)


权重: kata1-b40c256-s7850687232-d1904191439.bin.gz
显卡: 2060 SUPER
线程: 20:: EloDiff: +163    Visits/s: 822.99    nnEvals/s: 571.58    nnBatches/s: 57.82    avgBatchSize: 9.89(61.0secs)


权重: kata1-b40c256-s7850687232-d1904191439.bin.gz
显卡: 2060
线程: 16:: EloDiff: +111    Visits/s: 638.61    nnEvals/s: 464.32    nnBatches/s: 58.56    avgBatchSize: 7.93(78.5secs)


权重: kata1-b40c256-s7850687232-d1904191439.bin.gz
显卡: 1060
线程: 13:: EloDiff: +121    Visits/s: 345.62    nnEvals/s: 241.21    nnBatches/s: 37.38    avgBatchSize: 6.45(145.0secs)


权重: kata1-785.bin.gz
显卡: 2060
线程: 13:: EloDiff: +118    Visits/s: 535.95    nnEvals/s: 372.31    nnBatches/s: 57.75    avgBatchSize: 6.45(93.5secs)


权重: kata1bs790.bin.gz
显卡: 2080 Ti, 2070 SUPER, 2070 SUPER, 2080 Ti
线程: 58:: EloDiff: +447    Visits/s: 3085.61    nnEvals/s: 2304.23    nnBatches/s: 339.37    avgBatchSize: 6.79(16.4secs)


权重: kata1bs790.bin.gz
显卡: 2080 Ti, 2080 Ti, 2070 SUPER, 2080 Ti
线程: 64:: EloDiff: +473    Visits/s: 3382.32    nnEvals/s: 2567.71    nnBatches/s: 341.64    avgBatchSize: 7.52(15.0secs)
线程: 69:: EloDiff: +400    Visits/s: 3464.83    nnEvals/s: 2607.03    nnBatches/s: 353.87    avgBatchSize: 7.37(14.6secs)
线程: 89:: EloDiff: +466    Visits/s: 3610.82    nnEvals/s: 2751.05    nnBatches/s: 292.10    avgBatchSize: 9.42(14.1secs)
线程: 85:: EloDiff: +478    Visits/s: 3505.97    nnEvals/s: 2698.71    nnBatches/s: 295.36    avgBatchSize: 9.14(14.5secs)


权重: 60b_newAVG20210425.bin.gz
显卡: 2080 Ti, 2070 SUPER, 2070 SUPER, 2080 Ti
线程: 35:: EloDiff: +298    Visits/s: 1234.90    nnEvals/s: 977.11    nnBatches/s: 243.94    avgBatchSize: 4.01(16.5secs)


权重: 60b_newAVG20210425.bin.gz
显卡: 2080 Ti, 2080 Ti, 2070 SUPER, 2080 Ti
线程: 43:: EloDiff: +304    Visits/s: 1329.69    nnEvals/s: 1103.47    nnBatches/s: 246.14    avgBatchSize: 4.48(15.4secs)
线程: 44:: EloDiff: +329    Visits/s: 1388.00    nnEvals/s: 1118.15    nnBatches/s: 245.13    avgBatchSize: 4.56(14.7secs)


权重: kata1-b40c256-s7907049728-d1917596640.bin.gz
显卡: 1070 Ti
线程: 13:: EloDiff: +129    Visits/s: 324.73    nnEvals/s: 228.18    nnBatches/s: 29.11    avgBatchSize: 7.84(154.3secs)


权重: kata1-b40c256-s7907049728-d1917596640.bin.gz
显卡: 1050 Ti
线程: 8:: EloDiff: +59    Visits/s: 109.85    nnEvals/s: 82.36    nnBatches/s: 20.69    avgBatchSize: 3.98(292.0secs)


权重: kata1bs800.bin.gz
显卡: 2080 Ti, 2080 Ti, 2070 SUPER, 2080 Ti
线程: 75:: EloDiff: +472    Visits/s: 3361.82    nnEvals/s: 2573.31    nnBatches/s: 327.50    avgBatchSize: 7.86(15.1secs)
线程: 75:: EloDiff: +462    Visits/s: 3351.83    nnEvals/s: 2537.72    nnBatches/s: 314.11    avgBatchSize: 8.08(15.1secs)


权重: kata1bs804.bin.gz
显卡: 2080 Ti, 2080 Ti, 2070 SUPER, 2080 Ti
线程: 89:: EloDiff: +488    Visits/s: 3551.34    nnEvals/s: 2689.12    nnBatches/s: 284.43    avgBatchSize: 9.45(14.3secs)


权重: kata1bs809.bin.gz
显卡: 2080 Ti, 2080 Ti, 2070 SUPER, 2080 Ti
线程: 77:: EloDiff: +454    Visits/s: 3352.04    nnEvals/s: 2564.75    nnBatches/s: 301.72    avgBatchSize: 8.50(15.1secs)


权重: kata1-b40c256-s8003455744-d1941206379.bin.gz
显卡: 2060 SUPER
线程: 12:: EloDiff: +96    Visits/s: 759.91    nnEvals/s: 542.22    nnBatches/s: 84.51    avgBatchSize: 6.42(65.9secs)

<
最大线程数和核心线程数是线程池中的两个重要参数,它们在并发编程中起着关键的作用。 1. 最大线程数(Maximum Thread Count):最大线程数是指线程池中允许存在的最大线程数量。当有新的任务提交给线程池时,如果当前活动线程数小于最大线程数,线程池会创建新的线程来执行任务。如果当前活动线程数已达到最大线程数,并且任务队列也已满,则根据线程池的策略来处理无法执行的任务(如抛出异常或阻塞等)。 2. 核心线程数(Core Thread Count):核心线程数是指线程池中保持活动状态的最小线程数量。即使没有可执行的任务,核心线程也会一直存活。核心线程用于处理即时到达的任务,避免了频繁地创建和销毁线程的开销。当任务数量超过核心线程数时,线程池会根据任务队列和最大线程数来动态调整线程数量。 在Java中,可以使用`ThreadPoolExecutor`类来创建线程池,并通过其构造函数设置最大线程数和核心线程数。以下是一个示例: ```java import java.util.concurrent.ExecutorService; import java.util.concurrent.Executors; public class ThreadPoolExample { public static void main(String[] args) { int corePoolSize = 5; // 核心线程数 int maxPoolSize = 10; // 最大线程数 ExecutorService executorService = Executors.newFixedThreadPool(corePoolSize); // 或者使用 ThreadPoolExecutor 类来创建线程池 // ThreadPoolExecutor executor = new ThreadPoolExecutor(corePoolSize, maxPoolSize, 0L, TimeUnit.MILLISECONDS, new LinkedBlockingQueue<Runnable>()); // 向线程池提交任务 for (int i = 0; i < 20; i++) { executorService.execute(new RunnableTask()); } executorService.shutdown(); } static class RunnableTask implements Runnable { @Override public void run() { System.out.println("Executing task in thread: " + Thread.currentThread().getName()); // 执行具体的任务逻辑 } } } ``` 在上述示例中,我们使用`Executors.newFixedThreadPool()`方法创建一个固定大小的线程池,同时指定了核心线程数为`corePoolSize`。你还可以使用`ThreadPoolExecutor`类来创建线程池,可以更加灵活地设置各种参数。 请注意,最大线程数和核心线程数的具体取值应该根据应用程序的需求和系统资源进行调整,以避免过度消耗资源或导致线程饥饿等问题。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值