From 3f6ed46beac45cfc75b7d2300677c1623613ae9d Mon Sep 17 00:00:00 2001 From: Joan Fontanals Martinez Date: Mon, 7 Nov 2022 15:15:21 +0100 Subject: [PATCH] docs: document UUID support --- docs/fundamentals/flow/topologies.md | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/docs/fundamentals/flow/topologies.md b/docs/fundamentals/flow/topologies.md index 3eaa6d7cda626..0370ea153bf7e 100644 --- a/docs/fundamentals/flow/topologies.md +++ b/docs/fundamentals/flow/topologies.md @@ -147,7 +147,7 @@ You can restrict the visible devices in round-robin assignment using `CUDA_VISIB | 0 | 4 | -You can restrict the visible devices in round-robin assignment by assigning a list of devices ids `CUDA_VISIBLE_DEVICES=RR1,3`. This creates the following assignment: +You can restrict the visible devices in round-robin assignment by assigning a list of devices IDS `CUDA_VISIBLE_DEVICES=RR1,3`. This creates the following assignment: | GPU device | Replica ID | |------------|------------| @@ -157,6 +157,15 @@ You can restrict the visible devices in round-robin assignment by assigning a li | 3 | 3 | | 1 | 4 | +If your [CUDA driver](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/index.html#cuda-baremetal) allows to, you can also refer to GPUs by their UUID. For instance, you could assign a list of device UUIDs `CUDA_VISIBLE_DEVICES=RRGPU-0aaaaaaa-74d2-7297-d557-12771b6a79d5,GPU-0bbbbbbb-74d2-7297-d557-12771b6a79d5,GPU-0ccccccc-74d2-7297-d557-12771b6a79d5,GPU-0ddddddd-74d2-7297-d557-12771b6a79d5` + +| GPU device | Replica ID | +|------------|------------| +| GPU-0aaaaaaa-74d2-7297-d557-12771b6a79d5 | 0 | +| GPU-0bbbbbbb-74d2-7297-d557-12771b6a79d5 | 1 | +| GPU-0ccccccc-74d2-7297-d557-12771b6a79d5 | 2 | +| GPU-0ddddddd-74d2-7297-d557-12771b6a79d5 | 3 | +| GPU-0aaaaaaa-74d2-7297-d557-12771b6a79d5 | 4 | ## Distributed replicas