Running tensorflow without AVX on two xeon X5670

What works:
I followed tutorial and gpu in docker works (this passes)

docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark

I also recompiled Tensorflow (as I needed version without AVX) in docker container ( latest-gpu-py3 ) and installed it on ubuntu inside wsl2 it works on CPU but due to some issues with libraries it won’t work in Cuda (some random stuff is missing and this isn’t the problem).

The issue is that when I try to shamelessly use my tensorflow-2.1.0-cp36-cp36m-linux_x86_64.whl I have:

root@0d2d282f8f6c:/# python
Python 3.6.9 (default, Nov 7 2019, 10:44:02)
[GCC 8.3.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.

import tensorflow as tf
2020-07-05 16:52:22.113695: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1

tf.constant(3)
2020-07-05 16:52:28.949687: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2020-07-05 16:52:29.363319: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.364059: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1544] Found device 0 with properties:
pciBusID: 0000:03:00.0 name: GeForce GTX 1080 computeCapability: 6.1
coreClock: 1.8225GHz coreCount: 20 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 298.32GiB/s
2020-07-05 16:52:29.364190: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
2020-07-05 16:52:29.366146: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10
2020-07-05 16:52:29.367708: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10
2020-07-05 16:52:29.368058: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10
2020-07-05 16:52:29.370114: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10
2020-07-05 16:52:29.371511: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10
2020-07-05 16:52:29.376159: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-07-05 16:52:29.376833: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.378117: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.378422: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1686] Adding visible gpu devices: 0
2020-07-05 16:52:29.387851: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 2925995000 Hz
2020-07-05 16:52:29.388541: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x3bfe840 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-07-05 16:52:29.388564: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
2020-07-05 16:52:29.794702: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.795347: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x4318450 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-07-05 16:52:29.795409: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): GeForce GTX 1080, Compute Capability 6.1
2020-07-05 16:52:29.797755: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.798350: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1544] Found device 0 with properties:
pciBusID: 0000:03:00.0 name: GeForce GTX 1080 computeCapability: 6.1
coreClock: 1.8225GHz coreCount: 20 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 298.32GiB/s
2020-07-05 16:52:29.798526: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
2020-07-05 16:52:29.798642: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10
2020-07-05 16:52:29.798749: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10
2020-07-05 16:52:29.798862: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10
2020-07-05 16:52:29.798962: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10
2020-07-05 16:52:29.799025: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10
2020-07-05 16:52:29.799147: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-07-05 16:52:29.799766: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.800590: E tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:968] could not open file to read NUMA node: /sys/bus/pci/devices/0000:03:00.0/numa_node
Your kernel may have been built without NUMA support.
2020-07-05 16:52:29.801015: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1686] Adding visible gpu devices: 0
2020-07-05 16:52:29.801196: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1

Followed by nothing no error as if it’s stuck somewhere I’m suspecting it has to do with

Your kernel may have been built without NUMA support.

but I don’t understand what kernel doesn’t have this support and how to fix it.