Felix Abecassis

You likely need to take the latest version of TensorRT instead, yes. You will get better performance. You will probably need to make small edits to the Dockefile for that.

You have to edit the code in this case, to modify the `Classify` call.

> what GPU Allocated memory checks and "Zombie processes" checks do? Those are related, and it's related to reset too. If it's not possible to reset your card, you at...

I've added a wiki page on how to use MPS with Docker Compose: https://github.com/NVIDIA/nvidia-docker/wiki/MPS-(EXPERIMENTAL) You can look at the `docker-compose.yml` file for implementation details.

@azazhu are you running RHEL/Atomic's fork of Docker? If you do, you can just remove the `runtime:` lines and it should work fine. That's the `docker` package on RHEL/CentOS and...

Yes, that should work. But you can also containerize the MPS daemon, like in the Docker Compose example. I need to document the steps with the docker CLI too. >...

Yes, you can launch it inside a container or on the host. Both ways will work.

@GoodJoey not with the approach documented above, you would need a Volta GPU.

Another potential issue with the current naming is that individual tests might get mixed up if the test changes. For instance, with my parameters list above, let's say we have...

It doesn't seem to work: ``` [pid 32019] read(16, "\355[fUl\313\374\331\314*\246\256\237\16\246\331\225\370\356#>\\\240\224\365\215x>q-\247\253"..., 1048576) = 1048576 [pid 32019] write(17, "\355[fUl\313\374\331\314*\246\256\237\16\246\331\225\370\356#>\\\240\224\365\215x>q-\247\253"..., 1048576) = 1048576 [pid 32019] read(16, "\264\245IU#\17\306\322j\256\205\320\216\391\235\222)\344U\215\2667\269\\\216\276\325Qb"..., 1048576) = 494700 [pid 32019] read(16,...