"Attempting to perform BLAS operation using StreamExecutor without BLAS support" error occurs

my computer has only 1 GPU.

Below is what I get the result by entering someone's code

[name: "/device:CPU:0" device_type: "CPU" memory_limit: 268435456 locality {} incarnation: 16894043898758027805, name: "/device:GPU:0" device_type: "GPU" memory_limit: 10088284160 locality {bus_id: 1 links {}} incarnation: 17925533084010082620 physical_device_desc: "device: 0, name: GeForce RTX 3060, pci bus id: 0000:17:00.0, compute capability: 8.6"] 

I use jupyter notebook and I run 2 kernels now. (TensorFlow 2.6.0 and also installed CUDA and cuDNN as TensorFlow guide)

The first kernel is no problem to run my Sequential model from Keras.

But when I learn the same code in the second kernel, I got the error as below.

Attempting to perform BLAS operation using StreamExecutor without BLAS support [[node sequential_3/dense_21/MatMul (defined at \AppData\Local\Temp/ipykernel_14764/3692363323.py:1) ]] [Op:__inference_train_function_7682]

Function call stack: train_function

how can I learn multiple kernels without any problem and share them with only 1 GPU?

I am not familiar with TensorFlow 1.x.x version though.

I just solved this problem as below. This problem is because when keras run with gpu. It uses almost all vram. So i needed to give memory_limit for each notebook. Here is my code how i could solve it. You can just change memory_limit value.

gpus = tf.config.experimental.list_physical_devices('GPU') if gpus: try: tf.config.experimental.set_virtual_device_configuration( gpus[0],[tf.config.experimental.VirtualDeviceConfiguration(memory_limit=5120)]) except RuntimeError as e: print(e) 

2 Answers

For the benefit of community providing solution here

This problem is because when keras run with gpu, it uses almost all vram. So we needed to give memory_limit for each notebook as shown below

gpus = tf.config.experimental.list_physical_devices('GPU') if gpus: try: tf.config.experimental.set_virtual_device_configuration( gpus[0],[tf.config.experimental.VirtualDeviceConfiguration(memory_limit=5120)]) except RuntimeError as e: print(e) 

(Paraphrased from MCPMH)

I had this error when trying to run a python script when a Jupyter notebook was open. Killing the notebook kernel before running the script did the trick. It seems that only one program can use the GPU in the same time.