ComputeCpp error when running any tf code

[inferrna]

branch dev/amd_gpu
Got error when running any tf code

2018-08-07 16:06:02.058459: I ./tensorflow/core/common_runtime/sycl/sycl_device.h:70] Found following OpenCL devices:
2018-08-07 16:06:02.058534: I ./tensorflow/core/common_runtime/sycl/sycl_device.h:72] id: 0, type: GPU, name: Baffin, vendor: Advanced Micro Devices, Inc., profile: FULL_PROFILE
2018-08-07 16:06:02.058557: I ./tensorflow/core/common_runtime/sycl/sycl_device.h:72] id: 1, type: CPU, name:        Intel(R) Xeon(R) CPU E5-2665 0 @ 2.40GHz, vendor: Intel(R) Corporation, profile: FULL_PROFILE
terminate called after throwing an instance of 'cl::sycl::detail::exception_implementation<(cl::sycl::detail::exception_types)7, cl::sycl::detail::exception_implementation<(cl::sycl::detail::exception_types)6, cl::sycl::exception> >

Also look at this report
codeplaysoftware/computecpp-sdk#133
so it seems to be all correct with my OpenCL and computecpp installations.

build config

$ cat .tf_configure.bazelrc
build --action_env PYTHON_BIN_PATH="/usr/bin/python3"
build --action_env PYTHON_LIB_PATH="/usr/lib/python3/dist-packages"
build --python_path="/usr/bin/python3"
build:gcp --define with_gcp_support=true
build:hdfs --define with_hdfs_support=true
build:s3 --define with_s3_support=true
build:kafka --define with_kafka_support=true
build:xla --define with_xla_support=true
build:gdr --define with_gdr_support=true
build:verbs --define with_verbs_support=true
build --action_env TF_NEED_OPENCL_SYCL="1"
build --action_env TF_NEED_COMPUTECPP="1"
build:sycl --cxxopt=-DTENSORFLOW_SYCL_NO_HALF=1
build:sycl_asan --cxxopt=-DTENSORFLOW_SYCL_NO_HALF=1
build:sycl_arm --cxxopt=-DTENSORFLOW_SYCL_NO_HALF=1
build --action_env COMPUTECPP_TOOLKIT_PATH="/usr/local/computecpp"
build --action_env TF_SYCL_BITCODE_TARGET="spir64"
build --action_env TF_NEED_CUDA="0"
build --action_env TF_DOWNLOAD_CLANG="0"
build --define grpc_no_ares=true
build:opt --copt=-march=native
build:opt --host_copt=-march=native
build:opt --define with_default_optimizations=true
build --copt=-DGEMMLOWP_ALLOW_SLOW_SCALAR_FALLBACK
build --host_copt=-DGEMMLOWP_ALLOW_SLOW_SCALAR_FALLBACK

build command was

bazel build --config=opt --config=sycl --cxxopt=-std=c++11 --verbose_failures --logging=4 //tensorflow/tools/pip_package:build_pip_package

[lukeiwanski]

@inferrna thanks for the report.

I believe @DuncanMcBain is on top of it codeplaysoftware/computecpp-sdk#133 (comment)