GPU assignment

[digirak]

Describe the issue I have asked for two GPUs and the code clearly request so does the shell script. However I get the following error. 019-05-18 17:03:44.558783: I tensorflow/core/common_runtime/process_util.cc:71] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance. Traceback (most recent call last): File "/home/rnath/Code/DeepNet.py", line 62, in conv1d=get_model() File "/home/rnath/Code/DeepNet.py", line 58, in get_model model = keras.utils.multi_gpu_model(model, gpus=2) File "/home/rnath/miniconda3/lib/python3.7/site-packages/keras/utils/multi_gpu_utils.py", line 181, in multi_gpu_model available_devices)) ValueError: To call multi_gpu_model with gpus=2, we expect the following devices to be available: ['/cpu:0', '/gpu:0', '/gpu:1']. However this machine only has: ['/cpu:0', '/xla_cpu:0']. Try reducing gpus.

Context This is after the job has been assigned.

!/bin/bash

SBATCH --job-name=DeepSpec

SBATCH --time=24:10:00 # hh:mm:ss

SBATCH --output=output_val.txt

SBATCH --ntasks=3

SBATCH --gres=gpu:2

SBATCH --mem-per-cpu=5120 # 5GB

SBATCH --mail-user=rakesh.nath@uliege.be

SBATCH --mail-type=ALL

SBATCH --partition=gpu2080ti

SBATCH --comment=DeepLSpectra

python /home/rnath/Code/DeepNet.py

[JoeriHermans]

Are you using tensorflow-gpu / keras-gpu? Because your framework simply doesn't seem to detect the GPUs. Which it should.

Joeri

[JoeriHermans]

One thing you can do to check if you installed your software stack correctly is to ssh into one of the nodes, open a Python interpreter and do:

from tensorflow.python.client import device_lib
print(device_lib.list_local_devices())

[digirak]

Hi I just realized I hadn't installed keras-gpu. Thanks!