调用方式 | 并发数 | 每个并发多少个连接 | 平均相应时间(ms) | 总耗时(s) | tps(个/s) | cpu | mem | CPU NUM | MEM |
Rserve方式(机器学习) | 20 | 2000 | 96.80535ms | 3872s | 10.33000759 | cpu资源全部占用 | >9.4% | 8个cpu | 16333856k |
200 | 2000 | 108.81601ms | 43526s | 9.189824181 | cpu资源全部占用 | >52.5% | 8个cpu | 16333856k | |
rJava(机器学习) | 20 | 2000 | 165.2032ms | 6608s | 6.053151513 | >=100% | >4.5% | 8个cpu | 16333856k |
200 | 2000 | 160.7054475ms | 64282s | 6.222564422 | >=100% | >7.1% | 8个cpu | 16333856k | |
python模型(机器学习RandomForest) 服务器:192.168.162.181 | 50 | 2000 | 119.50599ms | 11950s | 8.367781397 | >=98% | >6.7 | 4个cpu | 8061500k |
200 | 2000 | 121.44424ms | 48577s | 8.234231611 | >=98% | >6.7 | 4个cpu | 8061500k | |
python(逻辑回归) | 200 | 2000 | 10.553225ms | 4221s | 94.75776362 | >96% | >0.9 | 4个cpu | 8061500k |
python(机器学习xgboost) | 50 | 2000 | 262.47999ms | 26247s | 3.809814226 | >268% | >0.9 | 4个cpu | 8061500k |
drool(逻辑回归) | 200 | 200 | 3.412475ms | 136s | 293.0424399 | >=97% | 7.30% | 4个cpu | 8061500k |
|
|
|
|
|
|
|
|
|
|
机器学习模型性能测试结果(R或者python),逻辑回归的测试结果没有添加
最新推荐文章于 2023-11-14 16:16:40 发布