FAIL: //tensorflow/core/grappler/clusters:single_machine_test

[lukeiwanski]

System Info

  Name:                      Hawaii
  Vendor:                    Advanced Micro Devices, Inc.
  Device OpenCL C version:           OpenCL C 2.0 
  Driver version:                1912.5 (VM)
  Profile:                   FULL_PROFILE
  Version:                   OpenCL 2.0 AMD-APP (1912.5)
  Extensions:                    cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_khr_gl_depth_images cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes

ComputeCpp 0.2.0

To reproduce

bazel test --config=sycl --local_test_jobs=4 -k --test_lang_filters=cc,py --action_env=LD_PRELOAD=/usr/lib/libOpenCL.so.1 --test_timeout 300,750,1200,3600 //tensorflow/core/grappler/clusters:single_machine_test

Error

[ RUN      ] SingleMachineTest.Queue
2017-06-12 15:55:49.124896: I ./tensorflow/core/grappler/clusters/single_machine.cc:269] Starting new session
2017-06-12 15:55:49.138891: I ./tensorflow/core/grappler/clusters/single_machine.cc:257] Cleaning up previous session
2017-06-12 15:55:49.142450: I ./tensorflow/core/grappler/clusters/single_machine.cc:269] Starting new session
2017-06-12 15:55:54.143757: F ./tensorflow/core/grappler/clusters/single_machine_test.cc:90] Non-OK-status: cluster_->Run(item.graph, item.feed, item.fetch, &metadata) status: Deadline exceeded: Failed to run the graph after 5 seconds, aborting
external/bazel_tools/tools/test/test-setup.sh: line 159: 19098 Aborted                 (core dumped) "${TEST_PATH}" "$@"
(END)

[jwlawson]

For me the failure is in SingleMachineTest.CostModel:

[ RUN      ] SingleMachineTest.CostModel
2017-06-14 10:43:59.108073: I ./tensorflow/core/grappler/clusters/single_machine.cc:257] Cleaning up previous session
2017-06-14 10:43:59.108701: I ./tensorflow/core/grappler/clusters/single_machine.cc:269] Starting new session
./tensorflow/core/grappler/clusters/single_machine_test.cc:72: Failure
Expected: (8) <= (node.output_info(0).size()), actual: 8 vs 0
[  FAILED  ] SingleMachineTest.CostModel (300 ms)

And the SingleMachineTest.Queue test passes. Tried both up to date eigen_mehdi and dev/eigen_mehdi at be13983.

[lukeiwanski]

Yes that's the current issue.

[jwlawson]

Fixed by #114

[lukeiwanski]

Closing