安装带有 SYCL 支持的 TensorFlow

Question

我正在尝试使用 gdb 来跟踪具有 Eigen SYCL 支持的 Tensorflow 操作内核实现。 但是，当我尝试安装.whl 软件包时，会弹出一些关于fglrx 的错误消息。

错误信息

Compiling /tmp/pip-1vfYDJ-build/tensorflow-1.0.1.data/purelib/tensorflow/contrib/cudnn_rnn/ops/gen_cudnn_rnn_ops.py ...
  File "/tmp/pip-1vfYDJ-build/tensorflow-1.0.1.data/purelib/tensorflow/contrib/cudnn_rnn/ops/gen_cudnn_rnn_ops.py", line 1
    Error: Fail to load fglrx kernel module!
             ^
SyntaxError: invalid syntax ....

配置

• CPU：英特尔酷睿 i7-4790
• GPU：GeForce GT 630/PCIe/SSE2
• 操作系统：ubuntu 15.04
• 司机：Nvidia binary driver nvidia-352

命令

这是我根据教程使用的命令。

https://www.codeplay.com/portal/03-30-17-setting-up-tensorflow-with-opencl-using-sycl

• 配置

  Please specify the location of python. [Default is /usr/local/bin/python]: 
  Please specify optimization flags to use during compilation [Default is -march=native]: 
  Do you wish to use jemalloc as the malloc implementation? (Linux only) [Y/n] 
  jemalloc enabled on Linux
  Do you wish to build TensorFlow with Google Cloud Platform support? [y/N] 
  No Google Cloud Platform support will be enabled for TensorFlow
  Do you wish to build TensorFlow with Hadoop File System support? [y/N] 
  No Hadoop File System support will be enabled for TensorFlow
  Do you wish to build TensorFlow with the XLA just-in-time compiler (experimental)? [y/N] 
  No XLA support will be enabled for TensorFlow
  Found possible Python library paths: /usr/local/lib/python2.7/dist-packages
  /usr/lib/python2.7/dist-packages
  Please input the desired Python library path to use.  Default is [/usr/local/lib/python2.7/dist-packages]
  Using python library path: /usr/local/lib/python2.7/dist-packages
  Do you wish to build TensorFlow with OpenCL support? [y/N] y
  OpenCL support will be enabled for TensorFlow
  Do you wish to build TensorFlow with CUDA support? [y/N] 
  No CUDA support will be enabled for TensorFlow
  Please specify which C++ compiler should be used as the host C++ compiler. [Default is /usr/bin/clang++-3.6]: 
  Please specify which C compiler should be used as the host C compiler. [Default is /usr/bin/clang-3.6]: 
  

• bazel 构建：

bazel build -c dbg --config=sycl //tensorflow/tools/pip_package:build_pip_package

• 构建 pip 包：

bazel-bin/tensorflow/tools/pip_package/build_pip_package

• 安装 pip 包

sudo pip 安装： /tmp/tensorflow_pkg/tensorflow-1.0.1-cp27-none-linux_x86_64.whl

请帮我解决问题。

Answer 1

您（目前）不能在英特尔 GPU 上将 SYCL 与 TensorFlow 结合使用。然而，它很快就会到来。您需要进行一些修复，然后它将正常工作。您需要等待新的 Intel OpenCL GPU 驱动程序，然后对 TensorFlow 进行一些兼容性提交，然后它才能在 Intel GPU 上运行。您可能还想再等一会以获得一些性能改进，因为我们一直更关注正确性，首先，性能会稍晚一些。

Answer 2

SYCL 变体目前需要支持 SPIR 或 SPIRV 的平台，例如 Intel 或 AMD。 NVIDIA 平台目前都不支持。