CODEWITHSUNDEEP
tensorflowmachine-learningkerasgpu
shogo
shogo

Reputation: 1

Although pc recognizes gpu, it uses a CPU in tensorflow-gpu

I am using tensorflow-gpu. I want to use GTX1070, but tensorflow-gpu uses my CPU. I don't know　what to do.

I use CUDA 9.0 and CUDNN 7.1.4. My tensorflow-gpu version is 1.9.

After running this command on the official website

>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()

2018-07-30 10:53:43.369025: I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 2018-07-30 10:53:43.829922: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:1392] Found device 0 with properties: name: GeForce GTX 1070 major: 6 minor: 1 memoryClockRate(GHz): 1.683 pciBusID: 0000:01:00.0 totalMemory: 8.00GiB freeMemory: 6.63GiB 2018-07-30 10:53:43.919043: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:1392] Found device 1 with properties: name: GeForce GTX 1050 major: 6 minor: 1 memoryClockRate(GHz): 1.455 pciBusID: 0000:05:00.0 totalMemory: 2.00GiB freeMemory: 1.60GiB 2018-07-30 10:53:43.926001: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:1456] Ignoring visible gpu device (device: 1, name: GeForce GTX 1050, pci bus id: 0000:05:00.0, compute capability: 6.1) with Cuda multiprocessor count: 5. The minimum required count is 8. You can adjust this requirement with the env var TF_MIN_GPU_MULTIPROCESSOR_COUNT. 2018-07-30 10:53:43.934810: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:1471] Adding visible gpu devices: 0 2018-07-30 10:53:44.761551: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:952] Device interconnect StreamExecutor with strength 1 edge matrix: 2018-07-30 10:53:44.765678: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:958] 0 1 2018-07-30 10:53:44.768363: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:971] 0: N N 2018-07-30 10:53:44.771773: I T:\src\github\tensorflow\tensorflow\core\common_runtime\gpu\gpu_device.cc:971] 1: N N 2018-07-30 10:53:44.774913: I T:\src\github\tensorflow\tensorflow\coenter code herere\common_runtime\gpu\gpu_device.cc:1084] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6395 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1070, pci bus id: 0000enter code here:01:00.0, compute capability: 6.1)

Upvotes: 0

Views: 1906

Answers (1)

Rob
Rob

Reputation: 563

As I can see from the log excerpt of your tensorflow engine - it uses GPU device 0

(/job:localhost/replica:0/task:0/device:GPU:0 with 6395 MB memory) -> physical 
GPU (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0, compute 
capability: 6.1)

But refuses to use your GeForce GTX 1050. This is possible due to the Environment variable TF_MIN_GPU_MULTIPROCESSOR_COUNT which is seems set to 8.

Try to set it to value of 5 as advised in your log earlier:

set TF_MIN_GPU_MULTIPROCESSOR_COUNT=5

If you want to be sure which device is used - initialize the session with 

sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))

You can read more on Using GPUs tensorflow docs page

Upvotes: 1

Related Questions