Skip to content

Instantly share code, notes, and snippets.

@myleott
Created February 3, 2020 17:00
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save myleott/fa49c10039c89b9472e6b0c59590b10b to your computer and use it in GitHub Desktop.
Save myleott/fa49c10039c89b9472e6b0c59590b10b to your computer and use it in GitHub Desktop.
Metric: CompileTime
TotalSamples: 2
Accumulator: 226ms137.620us
ValueRate: 439ms863.608us / second
Rate: 3.88139 / second
Percentiles: 1%=109ms635.741us; 5%=109ms635.741us; 10%=109ms635.741us; 20%=109ms635.741us; 50%=118ms501.879us; 80%=118ms501.879us; 90%=118ms501.879us; 95%=118ms501.879us; 99%=118ms501.879us
Metric: DeviceLockWait
TotalSamples: 101
Accumulator: 31s573ms487.494us
ValueRate: 754ms252.918us / second
Rate: 2.49169 / second
Percentiles: 1%=006.378us; 5%=300ms497.944us; 10%=302ms046.306us; 20%=306ms291.569us; 50%=310ms672.137us; 80%=313ms718.285us; 90%=313ms465.406us; 95%=314ms063.456us; 99%=316ms076.992us
Metric: ExecuteTime
TotalSamples: 99
Accumulator: 35s114ms639.771us
ValueRate: 963ms583.202us / second
Rate: 2.71392 / second
Percentiles: 1%=283ms535.438us; 5%=352ms274.296us; 10%=353ms595.082us; 20%=353ms052.365us; 50%=356ms650.115us; 80%=356ms445.762us; 90%=357ms812.043us; 95%=357ms966.990us; 99%=377ms816.817us
Metric: OutboundData
TotalSamples: 32
Accumulator: 919.44MB
ValueRate: 23.08MB / second
Rate: 0.8033 / second
Percentiles: 1%=2.00B; 5%=2.00B; 10%=8.00B; 20%=128.00KB; 50%=128.00KB; 80%=128.00KB; 90%=128.00KB; 95%=128.00KB; 99%=916.06MB
Metric: ReleaseDataHandlesTime
TotalSamples: 296
Accumulator: 01s183ms680.319us
ValueRate: 032ms422.675us / second
Rate: 8.11471 / second
Percentiles: 1%=001ms320.610us; 5%=002ms598.886us; 10%=002ms756.453us; 20%=002ms110.738us; 50%=003ms260.095us; 80%=006ms108.395us; 90%=007ms613.184us; 95%=007ms059.403us; 99%=010ms730.887us
Metric: TensorsGraphSize
TotalSamples: 100
Accumulator: 868605.00
ValueRate: 23620.53 / second
Rate: 2.71936 / second
Percentiles: 1%=8703.00; 5%=8703.00; 10%=8703.00; 20%=8703.00; 50%=8703.00; 80%=8703.00; 90%=8703.00; 95%=8703.00; 99%=8703.00
Metric: TransferToServerTime
TotalSamples: 32
Accumulator: 04s044ms554.888us
ValueRate: 102ms523.109us / second
Rate: 0.803436 / second
Percentiles: 1%=002ms708.557us; 5%=002ms966.467us; 10%=007ms305.074us; 20%=007ms449.305us; 50%=008ms851.804us; 80%=008ms288.711us; 90%=027ms512.333us; 95%=355ms673.686us; 99%=03s419ms971.958us
Counter: CachedCompile
Value: 98
Counter: CreateCompileHandles
Value: 2
Counter: CreateDataHandles
Value: 67771
Counter: CreateXlaTensor
Value: 307523
Counter: DestroyDataHandles
Value: 67080
Counter: DestroyXlaTensor
Value: 306835
Counter: MarkStep
Value: 101
Counter: ReleaseDataHandles
Value: 67080
Counter: SyncTensorsToData
Value: 339
Counter: UncachedCompile
Value: 2
Counter: XRTAllocateFromTensor_Empty
Value: 349
Counter: XrtCompile_Empty
Value: 32
Counter: XrtExecuteChained_Empty
Value: 32
Counter: XrtExecute_Empty
Value: 32
Counter: XrtRead_Empty
Value: 32
Counter: XrtReleaseAllocationHandle_Empty
Value: 32
Counter: XrtReleaseCompileHandle_Empty
Value: 32
Counter: XrtSessionCount
Value: 4
Counter: XrtSubTuple_Empty
Value: 32
Counter: xla::_log_softmax
Value: 101
Counter: xla::_log_softmax_backward_data
Value: 101
Counter: xla::_unsafe_view
Value: 12221
Counter: xla::add
Value: 9696
Counter: xla::add_
Value: 80360
Counter: xla::addcmul
Value: 4848
Counter: xla::as_strided
Value: 339
Counter: xla::bernoulli_
Value: 4848
Counter: xla::copy_
Value: 339
Counter: xla::div_
Value: 4848
Counter: xla::embedding
Value: 101
Counter: xla::embedding_dense_backward
Value: 101
Counter: xla::empty
Value: 5288
Counter: xla::empty_strided
Value: 339
Counter: xla::fill_
Value: 101
Counter: xla::index_select
Value: 101
Counter: xla::mm
Value: 36663
Counter: xla::mul
Value: 24240
Counter: xla::native_batch_norm
Value: 4848
Counter: xla::native_batch_norm_backward
Value: 4848
Counter: xla::native_layer_norm
Value: 4848
Counter: xla::native_layer_norm_backward
Value: 4848
Counter: xla::nll_loss_backward
Value: 101
Counter: xla::nll_loss_forward
Value: 101
Counter: xla::relu
Value: 2424
Counter: xla::sub
Value: 4848
Counter: xla::sum
Value: 21917
Counter: xla::t
Value: 48884
Counter: xla::threshold_backward
Value: 2424
Counter: xla::view
Value: 110090
Metric: XrtAllocateFromTensor
TotalSamples: 39046
Accumulator: 12m34s563ms326.838us
Mean: 028ms718.313us
StdDev: 037ms906.940us
Rate: 3.213 / second
Percentiles: 25%=006ms617.655us; 50%=010ms435.398us; 80%=058ms856.395us; 90%=101ms174.391us; 95%=106ms521.087us; 99%=112ms454.181us
Metric: XrtCompile
TotalSamples: 148
Accumulator: 24m32s865ms754.480us
Mean: 10s540ms626.719us
StdDev: 13s024ms674.761us
Rate: 0.00297496 / second
Percentiles: 25%=066ms689.293us; 50%=080ms984.058us; 80%=24s074ms154.801us; 90%=32s903ms002.383us; 95%=35s896ms246.678us; 99%=35s960ms566.343us
Metric: XrtExecute
TotalSamples: 9893
Accumulator: 49m39s225ms285.246us
Mean: 316ms581.172us
StdDev: 661ms575.966us
Rate: 0.211111 / second
Percentiles: 25%=180ms198.086us; 50%=211ms358.858us; 80%=353ms791.811us; 90%=354ms834.165us; 95%=354ms249.498us; 99%=03s855ms207.106us
Metric: XrtReleaseAllocation
TotalSamples: 23489
Accumulator: 01m09s255ms598.255us
Mean: 003ms551.585us
StdDev: 002ms975.566us
Rate: 2.59503 / second
Percentiles: 25%=942.872us; 50%=001ms497.611us; 80%=005ms050.236us; 90%=005ms442.076us; 95%=006ms731.992us; 99%=006ms124.971us
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment