Skip to content

Instantly share code, notes, and snippets.

@kazimuth
Last active September 29, 2019 20:54
Show Gist options
  • Save kazimuth/6ffa0d6035fd38f1bef58dc512c8571b to your computer and use it in GitHub Desktop.
Save kazimuth/6ffa0d6035fd38f1bef58dc512c8571b to your computer and use it in GitHub Desktop.
import Glibc
setenv("SWIFT_TENSORFLOW_ENABLE_DEBUG_LOGGING", "true", 1)
let s = "grpc://\(String(cString: getenv("COLAB_TPU_ADDR")!))"
setenv("SWIFT_TENSORFLOW_SERVER_ADDRESS", s, 1)
print("SWIFT_TENSORFLOW_SERVER_ADDRESS:", String(cString: getenv("SWIFT_TENSORFLOW_SERVER_ADDRESS")!))
import TensorFlow
func runTestComputation() -> Tensor<Float> {
// Create two 1000x1000 tensors sampled from a random normal distribution.
let x = Tensor<Float>(randomNormal: [1000, 1000])
let w = Tensor<Float>(randomNormal: [1000, 1000])
// Multiply them together and return the result.
return (w * x).mean()
}
print("~~RUNNING CPU~~")
let x = withDevice(.cpu, 0) {
runTestComputation()
}
print("CPU MEAN:", x)
print()
print("~~RUNNING TPU~~")
let y = withDevice(.tpu, 0) {
runTestComputation()
}
print("TPU MEAN:", y)
SWIFT_TENSORFLOW_SERVER_ADDRESS: grpc://10.126.146.18:8470
~~RUNNING CPU~~
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:89] Turning on debug logging from env.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:111] Env var SWIFT_TENSORFLOW_SERVER_ADDRESS has value grpc://10.126.146.18:8470.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:142] Setting TF server address to grpc://10.126.146.18:8470 from env.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:151] Setting TF_EAGER_REMOTE_USE_SEND_TENSOR_RPC to 1
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:195] Initializing global context.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:249] Allowing growth for GPU memory allocator.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:268] Setting up the server def to cluster {
job {
name: "localhost"
tasks {
key: 0
value: "127.0.0.1:0"
}
tasks {
key: 1
value: "10.126.146.18:8470"
}
}
}
job_name: "localhost"
task_index: 0
protocol: "grpc"...
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:283] There are 13 devices.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 0 has type CPU and name /job:localhost/replica:0/task:0/device:CPU:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 1 has type XLA_CPU and name /job:localhost/replica:0/task:0/device:XLA_CPU:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 2 has type CPU and name /job:localhost/replica:0/task:1/device:CPU:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 3 has type XLA_CPU and name /job:localhost/replica:0/task:1/device:XLA_CPU:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 4 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 5 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:1.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 6 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:2.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 7 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:3.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 8 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:4.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 9 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:5.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 10 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:6.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 11 has type TPU and name /job:localhost/replica:0/task:1/device:TPU:7.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:291] Device 12 has type TPU_SYSTEM and name /job:localhost/replica:0/task:1/device:TPU_SYSTEM:0.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:49] De-initializing TensorHandle.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:52] Returning from deinit of TensorHandle.
CPU MEAN: [/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Tensor.swift:160] Returning a host copy of array.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/TensorHandle.swift:177] Calling makeHostCopy() with c handle 0x000000000350a0d8
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:481] # of dims is 0
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:482] Returning a shaped array.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:461] Initializing ShapedArray from CTensor.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:467] The shape is [].
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:74] Initializing TensorBuffer with a cTensor of 1 elements.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:470] Done initializing ShapedArray from CTensor.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:53] De-initializing tensor buffer.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:58] Deleting underlying tensor.
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/ShapedArray.swift:61] Returning from deinit of TensorBuffer.
0.00011846422
~~RUNNING TPU~~
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:424] Calling _TFCEagerExecute() over:
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:429] [Run with TF_CPP_MIN_LOG_LEVEL=0 to have TFEOps printed out]
[/swift-base/tensorflow-swift-apis/Sources/TensorFlow/Core/Runtime.swift:432] Executing eager op 0x0000000003055080.
Fatal error: Failed copying input tensor from /job:localhost/replica:0/task:0/device:CPU:0 to /job:localhost/replica:0/task:1/device:CPU:0 in order to run StatelessRandomNormal: Unable to find a context_id matching the specified one (-3734417892677293884). Perhaps the worker was restarted, or the context was GC'd?
Additional GRPC error information:
{"created":"@1569626525.990149906","description":"Error received from peer","file":"external/grpc/src/core/lib/surface/call.cc","file_line":1039,"grpc_message":"Unable to find a context_id matching the specified one (-3734417892677293884). Perhaps the worker was restarted, or the context was GC'd?","grpc_status":3}: file /swift-base/tensorflow-swift-apis/Sources/TensorFlow/Bindings/EagerExecution.swift, line 300
Current stack trace:
0 libswiftCore.so 0x00007f9ac150f940 swift_reportError + 50
1 libswiftCore.so 0x00007f9ac1580cf0 _swift_stdlib_reportFatalErrorInFile + 115
2 libswiftCore.so 0x00007f9ac14aab48 <unavailable> + 3722056
3 libswiftCore.so 0x00007f9ac14aacd7 <unavailable> + 3722455
4 libswiftCore.so 0x00007f9ac127d4e8 <unavailable> + 1438952
5 libswiftCore.so 0x00007f9ac147e5ce <unavailable> + 3540430
6 libswiftCore.so 0x00007f9ac127cc09 <unavailable> + 1436681
7 libswiftTensorFlow.so 0x00007f9abe70af50 <unavailable> + 2899792
8 libswiftTensorFlow.so 0x00007f9abe55ed10 checkOk(_:file:line:) + 434
9 libswiftTensorFlow.so 0x00007f9abe565ce0 TFE_Op.evaluateUnsafe() + 506
10 libswiftTensorFlow.so 0x00007f9abe5663d0 TFE_Op.execute<A>(_:) + 132
11 libswiftTensorFlow.so 0x00007f9abe56f074 <unavailable> + 1212532
12 libswiftTensorFlow.so 0x00007f9abe69d810 static Raw.statelessRandomNormal<A, B, C>(shape:seed:) + 1320
13 libswiftTensorFlow.so 0x00007f9abe742930 Tensor<>.init(randomNormal:mean:standardDeviation:seed:) + 944
17 libswiftTensorFlow.so 0x00007f9abe6d30c1 <unavailable> + 2670785
18 libswiftTensorFlow.so 0x00007f9abe6d2cf6 <unavailable> + 2669814
19 libswiftTensorFlow.so 0x00007f9abe6d2780 withDevice<A>(_:_:perform:) + 169
21 repl_swift 0x0000000000400490 <unavailable> + 1168
Current stack trace:
frame #9: 0x00007f9ac3093a2d $__lldb_expr16`runTestComputation() at <Cell 1>:11:25
frame #10: 0x00007f9ac3093e09 $__lldb_expr16`specialized thunk for @callee_guaranteed () -> (@owned Tensor<Float>, @error @owned Error) [inlined] closure #1 () -> TensorFlow.Tensor<Swift.Float> in __lldb_expr_15 at <Cell 1>:19:4
frame #16: 0x00007f9ac3093646 $__lldb_expr16`main at <Cell 1>:24:9
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment