# RayOutOfMemoryError **URL:** \ **Category:** RLlib\ **Created:** [May 24, 2021, 10:10am UTC](https://discuss.ray.io/t/rayoutofmemoryerror/2268 "2021-05-24T10:10:16Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![Clement\_Collgon](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/clement_collgon/32/353_2.png) [@Clement\_Collgon](https://discuss.ray.io/u/Clement_Collgon)\ **Post date:** [May 24, 2021, 10:10am UTC](https://discuss.ray.io/t/rayoutofmemoryerror/2268/1 "2021-05-24T10:10:16Z")
Hello, I’m using ray RLLIB. It works perfectly well on all computers of my office, however, one with a bit more power gives us problem. When we launch runs on it, the memory will rise until it reaches maximum ( no dump of memory it seems ) and kill all active process one after another. Error : ray.\_private.memory\_monitor.RayOutOfMemoryError: More than 95% of the memory on node zeta-z1 is used (59.52 / 62.49 GB). The top 10 memory consumers are: PID MEM COMMAND 553077 1.76GiB ray::ImplicitFunc.train\_buffered() 553084 1.73GiB ray::ImplicitFunc.train\_buffered() 553083 1.73GiB ray::ImplicitFunc.train\_buffered() 553074 1.71GiB ray::ImplicitFunc.train\_buffered() 553086 1.52GiB ray::ImplicitFunc.train\_buffered() 553085 1.5GiB ray::ImplicitFunc.train\_buffered() 553131 1.49GiB ray::ImplicitFunc.train\_buffered() 553082 1.48GiB ray::ImplicitFunc.train\_buffered() 553087 1.46GiB ray::ImplicitFunc.train\_buffered() 553071 1.46GiB ray::ImplicitFunc.train\_buffered() Tensorboard : ![image](https://us1.discourse-cdn.com/flex020/uploads/ray/original/2X/b/b0d67edce6dec6708edb4afe60adffeaa36bddbc.png) Thanks in advance ! ---
**Author:** ![sven1977](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/sven1977/32/53_2.png) [@sven1977](https://discuss.ray.io/u/sven1977)\ **Post date:** [May 24, 2021, 10:18am UTC](https://discuss.ray.io/t/rayoutofmemoryerror/2268/2 "2021-05-24T10:18:10Z")
Hey @Clement_Collgon , can you give us more information on what Ray version, algorithm, RLlib config you are running with? Hard to diagnose from this distance w/o more information. Thanks. ---
**Author:** ![Clement\_Collgon](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/clement_collgon/32/353_2.png) [@Clement\_Collgon](https://discuss.ray.io/u/Clement_Collgon)\ **Post date:** [May 24, 2021, 10:24am UTC](https://discuss.ray.io/t/rayoutofmemoryerror/2268/3 "2021-05-24T10:24:31Z")
Dear Sven, the ray version 1.3.0, the RLLIB config is : “rllib”: { “grid\_search”: { “\_fake\_gpus”: false, “actor\_hidden\_activation”: “relu”, “actor\_hiddens”: [ 512, 512 ], “actor\_lr”: 0.0001, “batch\_mode”: “truncate\_episodes”, “buffer\_size”: 300000, “callbacks”: “\”, “clip\_actions”: true, “clip\_rewards”: null, “collect\_metrics\_timeout”: 180, “compress\_observations”: false, “create\_env\_on\_driver”: false, “critic\_hidden\_activation”: “relu”, “critic\_hiddens”: [ 512, 512 ], “critic\_lr”: 0.001, “custom\_eval\_function”: null, “custom\_resources\_per\_worker”: {}, “eager\_tracing”: false, “env”: null, “env\_config”: {}, “evaluation\_config”: { “env\_config”: {}, “explore”: false }, “evaluation\_interval”: null, “evaluation\_num\_episodes”: 10, “evaluation\_num\_workers”: 0, “exploration\_config”: { “final\_scale”: 1.0, “initial\_scale”: 1.0, “stddev”: 0.12, “type”: “GaussianNoise” }, “explore”: true, “extra\_python\_environs\_for\_driver”: {}, “extra\_python\_environs\_for\_worker”: {}, “fake\_sampler”: false, “final\_prioritized\_replay\_beta”: 0.4, “framework”: “tf”, “gamma”: 0.9, “grad\_clip”: null, “horizon”: null, “huber\_threshold”: 1.0, “ignore\_worker\_failures”: false, “in\_evaluation”: false, “input”: “sampler”, “input\_evaluation”: [ “is”, “wis” ], “l2\_reg”: 0, “learning\_starts”: 1500, “local\_tf\_session\_args”: { “inter\_op\_parallelism\_threads”: 8, “intra\_op\_parallelism\_threads”: 8 }, “log\_level”: “WARN”, “log\_sys\_usage”: true, “logger\_config”: null, “lr”: 0.0001, “metrics\_smoothing\_episodes”: 100, “min\_iter\_time\_s”: 1, “model”: { “\_time\_major”: false, “conv\_activation”: “relu”, “conv\_filters”: null, “custom\_action\_dist”: null, “custom\_model”: null, “custom\_model\_config”: {}, “custom\_preprocessor”: null, “dim”: 84, “fcnet\_activation”: “tanh”, “fcnet\_hiddens”: [ 256, 256 ], “framestack”: true, “free\_log\_std”: false, “grayscale”: false, “lstm\_cell\_size”: 256, “lstm\_use\_prev\_action”: false, “lstm\_use\_prev\_action\_reward”: -1, “lstm\_use\_prev\_reward”: false, “max\_seq\_len”: 20, “no\_final\_linear”: false, “use\_lstm”: false, “vf\_share\_layers”: true, “zero\_mean”: true }, “monitor”: -1, “multiagent”: { “count\_steps\_by”: “env\_steps”, “observation\_fn”: null, “policies”: {}, “policies\_to\_train”: null, “policy\_mapping\_fn”: null, “replay\_mode”: “independent” }, “n\_step”: 1, “no\_done\_at\_end”: false, “normalize\_actions”: false, “num\_cpus\_for\_driver”: 1, “num\_cpus\_per\_worker”: 0, “num\_envs\_per\_worker”: 1, “num\_gpus”: 0, “num\_gpus\_per\_worker”: 0, “num\_workers”: 10, “observation\_filter”: “NoFilter”, “optimizer”: {}, “output”: null, “output\_compress\_columns”: [ “obs”, “new\_obs” ], “output\_max\_file\_size”: 67108864, “placement\_strategy”: “PACK”, “policy\_delay”: 2, “postprocess\_inputs”: false, “preprocessor\_pref”: “deepmind”, “prioritized\_replay”: false, “prioritized\_replay\_alpha”: 0.6, “prioritized\_replay\_beta”: 0.4, “prioritized\_replay\_beta\_annealing\_timesteps”: 20000, “prioritized\_replay\_eps”: 1e-06, “record\_env”: false, “remote\_env\_batch\_wait\_ms”: 0, “remote\_worker\_envs”: false, “render\_env”: false, “rollout\_fragment\_length”: 1, “sample\_async”: false, “sample\_collector”: “\”, “seed”: null, “shuffle\_buffer\_size”: 0, “simple\_optimizer”: -1, “smooth\_target\_policy”: false, “soft\_horizon”: false, “synchronize\_filters”: true, “target\_network\_update\_freq”: 0, “target\_noise”: 0.12, “target\_noise\_clip”: 1.0, “tau”: 0.001, “tf\_session\_args”: { “allow\_soft\_placement”: true, “device\_count”: { “CPU”: 1 }, “gpu\_options”: { “allow\_growth”: true }, “inter\_op\_parallelism\_threads”: 2, “intra\_op\_parallelism\_threads”: 2, “log\_device\_placement”: false }, “timesteps\_per\_iteration”: 1000, “train\_batch\_size”: 512, “training\_intensity”: null, “twin\_q”: true, “use\_huber”: false, “use\_state\_preprocessor”: false, “worker\_side\_prioritization”: false }, “num\_samples”: 1, “policies\_config”: {}, “ressources\_GS”: { “cpu”: 2, “gpu”: 0 }, “trainer”: “DDPG” }, with the DDPG algorithm. The processor model name : Intel(R) Core™ i9-10920X CPU @ 3.50GHz The memory available : 64GB. Thanks a lot