News for hackers. Stuff that matters.
Re:Introducing Amazon EC2 P3 Instances by SloopJon on Thursday October 26, 2017 @02:22PM (#15561573)
This post states, "In order to take full advantage of the NVIDIA Tesla V100 GPUs and the Tensor cores, you will need to use CUDA 9 and cuDNN7." What version of TensorFlow does it use? From what I can tell, TensorFlow doesn't fully support the latest versions yet.
Re:Introducing Amazon EC2 P3 Instances by sumt on Thursday October 26, 2017 @07:22PM (#15564135)
You can use the new AWS Deep Learning AMI which has a version of TensorFlow enhanced for CUDA 9 and Volta support https://aws.amazon.com/blogs/ai/announcing-new-aws-deep-lear...
Re:Introducing Amazon EC2 P3 Instances by sipherhex on Friday October 27, 2017 @03:40AM (#15566422)
Chris from NV here. You can also get a full compliment of DL framework containers, as well as CUDA 9/CuDNN 7/NCCL 2 base container, optimized for Volta by NVIDIA via this AMI https://aws.amazon.com/marketplace/pp/B076K31M1S?qid=1509089...
Re:Introducing Amazon EC2 P3 Instances by dharma1 on Tuesday October 31, 2017 @05:31AM (#15591918)
Are there any frameworks yet supporting the advertised 120 TFLOPs mixed precision training?
Re:Introducing Amazon EC2 P3 Instances by puzzle on Thursday October 26, 2017 @02:35PM (#15561696)
TF 1.4 does, but you need to build it yourself. RC1 is out:

https://github.com/tensorflow/tensorflow/releases/tag/v1.4.0...

  All our prebuilt binaries have been built with CUDA 8 and cuDNN 6.
  We anticipate releasing TensorFlow 1.5 with CUDA 9 and cuDNN 7.