Commit 6d2cf8e7 authored by Michael Krause's avatar Michael Krause 🎉
Browse files

Update GPU groups/names

parent 1a44936d
Pipeline #8657 passed with stages
in 35 seconds
...@@ -6,37 +6,38 @@ flavors of Nvidia GPUs attached to them to be used with CUDA-enabled code. ...@@ -6,37 +6,38 @@ flavors of Nvidia GPUs attached to them to be used with CUDA-enabled code.
Right now we have these nodes available: Right now we have these nodes available:
======== =============== ====== ===== ========= ======== =============== ====== ===== ========= ============
Nodename GPU Type Memory Count Partition Nodename GPU Type Memory Count Partition Architecture
======== =============== ====== ===== ========= ======== =============== ====== ===== ========= ============
gpu-1 GTX 1080 TI 12 GB 2 test gpu-1 GTX 1080 TI 12 GB 2 test Pascal
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-2 GTX 1080 8 GB 3 gpu gpu-2 GTX 1080 8 GB 3 gpu Pascal
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-3 GTX 1080 8 GB 3 gpu gpu-3 GTX 1080 8 GB 3 gpu Pascal
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-4 Quadro RTX 5000 16 GB 4 gpu gpu-4 Quadro RTX 5000 16 GB 4 gpu Turing
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-5 Quadro RTX 5000 16 GB 4 gpu gpu-5 Quadro RTX 5000 16 GB 4 gpu Turing
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-6 Quadro RTX 5000 16 GB 4 gpu gpu-6 Quadro RTX 5000 16 GB 4 gpu Turing
-------- --------------- ------ ----- --------- -------- --------------- ------ ----- --------- ------------
gpu-7 Quadro RTX 5000 16 GB 4 gpu gpu-7 Quadro RTX 5000 16 GB 4 gpu Turing
======== =============== ====== ===== ========= ======== =============== ====== ===== ========= ============
Both the 12GB 1080 TI and the 8GB 1080 are grouped under the name **1080**. The Both the 12GB 1080 TI and the 8GB 1080 are grouped under the name **pascal**. The
short name for the more powerful Quadro cards is **rtx5k**. short name for the more powerful Quadro cards is **turing**.
To request any GPU, you can use ``-p gpu --gres gpu:1`` or ``-p test --gres To request any GPU, you can use ``-p gpu --gres gpu:1`` or ``-p test --gres
gpu:1`` if you want to test things. The ``gres`` parameter is very flexible and gpu:1`` if you want to test things. The ``gres`` parameter is very flexible and
allows to request the GPU group (**1080** or **rtx5k**). allows to request the GPU group/architecture (**pascal** or **turing**).
For example, to request 2 Geforce 1080, use ``--gres gpu:1080:2``. This will For example, to request 2 Geforce 1080, use ``--gres gpu:pascal:2``. This will
effectively hide all other GPUs and grants exclusive usage of the devices. effectively hide all other GPUs and grants exclusive usage of the devices.
You can use the `nvidia-smi` tool in an interactive job or the node-specific You can use the `nvidia-smi` tool in an interactive job or the node-specific
charts to get an idea of the device's utilization. charts to get an idea of the device's utilization.
Any code that supports CUDA up to version 10.1 should just work out of the box, that includes python's pygpu or Matlab's gpu-enabled libraries. Any code that supports CUDA up to version 10.1 should just work out of the box,
that includes python's pygpu or Matlab's gpu-enabled libraries.
.. note:: .. note::
...@@ -44,12 +45,15 @@ Any code that supports CUDA up to version 10.1 should just work out of the box, ...@@ -44,12 +45,15 @@ Any code that supports CUDA up to version 10.1 should just work out of the box,
container**. You have to pass the ``--nv`` flag to any container**. You have to pass the ``--nv`` flag to any
singylarity calls, however. singylarity calls, however.
Example: Request an interactive job (srun --pty) with 4 cores, 8gb of memory and a single card from the rtx5k group. Instead of ``/bin/bash`` we use the shell from a singularity container and tell singularity to prepare an nvidia environment ``singularity shell --nv``: Example: Request an interactive job (``srun --pty``) with 4 cores, 8 GB of
memory and a single card from the *turing* group. Instead of ``/bin/bash`` we use
the shell from a singularity container and tell singularity to prepare an
Nvidia environment with ``singularity shell --nv``:
.. code:: .. code::
srun --pty -p gpu --gres gpu:rtx5k:1 -c 4 --mem 8gb \ srun --pty -p gpu --gres gpu:turing:1 -c 4 --mem 8gb \
singularity shell --nv /data/container/unofficial/fsl/fsl-6.0.3.sif singularity shell --nv /data/container/unofficial/fsl/fsl-6.0.4.sif
Singularity> hostname Singularity> hostname
gpu-4 gpu-4
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment